Power Query for Analysts: Difference between revisions

From Training Material
Jump to navigation Jump to search
Gpruszczynski (talk | contribs)
 
(87 intermediate revisions by 2 users not shown)
Line 1: Line 1:
== Moduł 1: Wprowadzenie do Power Query i podstawowe transformacje ==
{{Cat|Power Query|001}}
== Module 1: Introduction to Power Query and Basic Transformations ==


=== Cel ===
=== Objective ===
W tym ćwiczeniu nauczysz się:
In this exercise, you will learn to:


Importować dane z pliku CSV do Power Query.
* Import data from a CSV file into Power Query.
Sprawdzać i dostosowywać typy danych, koncentrując się na konwersji daty zapisanej jako tekst na rzeczywisty typ daty.
* Review and adjust data types, focusing on converting a date stored as text into a proper date type.
Zastosować podstawowe filtrowanie danych.
* Apply basic data filtering.
Korzystać z pomocy zewnętrznej (np. ChatGPT) w celu uzyskania wskazówek dotyczących tworzenia niestandardowego kodu M, bez bezpośredniego kopiowania rozwiązań.
* Use external help (e.g., ChatGPT) to get guidance on creating custom M code, without directly copying solutions.
=== Dostarczone dane ===
Otrzymujesz do pobrania plik CSV o nazwie
📂[[PQsales.csv]]
który zawiera dane zamówień sprzedaży.


=== Instrukcje ===
=== Files Provided ===
You can download the CSV file 
📂[[Media:PQsales.csv]] 
which contains sales order data.


* Krok 1: Importuj dane
=== Instructions ===


#Otwórz Power Query w Excelu lub Power BI.
* Step 1: Import the data
#Zaimportuj dane z pliku sales.csv.
#Zauważ, że kolumna OrderDate została zaimportowana jako tekst ze względu na swój format (dd/MM/yyyy).
* Krok 2: Sprawdź i przekonwertuj typy danych


#Sprawdź, czy każda kolumna ma poprawny typ danych.
# Open Power Query in Excel or Power BI.
#Ręcznie przekształć kolumnę OrderDate z tekstu na typ daty.
# Import the data from the file sales.csv.
#Wskazówka: Jeśli napotkasz trudności, możesz zapytać ChatGPT o wskazówki, jak napisać funkcję M do konwersji tekstu na datę.
# Note that the *OrderDate* column was imported as text due to its format (dd/MM/yyyy).
* Krok 3: Zastosuj podstawowe filtrowanie


#Filtruj zestaw danych, tak aby pozostały tylko wiersze, w których koszt jest większy niż 200.
* Step 2: Check and convert data types
#Propozycja: Użyj interfejsu graficznego Power Query lub napisz prosty skrypt M, aby zastosować filtr. Jeśli to konieczne, skonsultuj się z ChatGPT w celu uzyskania pomysłów na implementację tego filtru.
* Krok 4: Przejrzyj i zapisz swoją pracę


#Potwierdź, że transformacje zostały poprawnie zastosowane, przeglądając podgląd danych.
# Verify that each column has the correct data type.
#Zapisz zapytanie i udokumentuj kroki, które podjąłeś.
# Manually transform the *OrderDate* column from text to date type.
# Tip: If you encounter difficulties, you can ask ChatGPT for hints on how to write an M function to convert text to date.


=== Zadanie ===
* Step 3: Apply basic filtering


Wykonaj kroki opisane powyżej w Power Query.
# Filter the dataset to keep only rows where the cost is greater than 200.
Eksperymentuj z dostępnymi opcjami transformacji i staraj się zrozumieć, jak każdy krok wpływa na Twoje dane.
# Suggestion: Use Power Query’s graphical interface or write a simple M script to apply the filter. If needed, consult ChatGPT for implementation ideas.
Używaj ChatGPT w celu uzyskania wskazówek lub rozwiązywania problemów, ale unikaj kopiowania kompletnych rozwiązań dosłownie.


== Moduł 2: Łączenie i scalanie danych z wielu źródeł ==
* Step 4: Review and save your work


=== Cel ===
# Confirm that the transformations were applied correctly by reviewing the data preview.
# Save the query and document the steps you have taken.


W tym ćwiczeniu nauczysz się:
=== Task ===


Importować dane z wielu plików CSV do Power Query. Scalić (połączyć) dane z różnych źródeł na podstawie wspólnego klucza. Użyć lewego zewnętrznego łączenia (Left Outer Join) do dodania szczegółów klienta do zamówień sprzedaży. Korzystać z pomocy zewnętrznej (np. ChatGPT) w celu uzyskania wskazówek dotyczących pisania niestandardowego kodu M, bez kopiowania kompletnych rozwiązań.
Perform the steps described above in Power Query.
=== Dostarczone dane ===
Experiment with available transformation options and try to understand how each step affects your data.
Otrzymujesz do pobrania dwa pliki CSV:
Use ChatGPT for hints or troubleshooting, but avoid copying complete solutions verbatim.


📂 [[Plik:PQ_sales2.csv]] – Zawiera dane zamówień sprzedaży:
== Module 2: Combining and Merging Data from Multiple Sources ==


📂 [[Plik:PQ_customers.csv]] – Zawiera informacje o klientach:
=== Objective ===


=== Instrukcje ===
In this exercise, you will learn to:


*Krok 1: Importuj dane
* Import data from multiple CSV files into Power Query.
Otwórz Power Query w Excelu lub Power BI. Zaimportuj dane z obu plików PQ_sales.csv i PQ_customers.csv. Sprawdź, czy oba zapytania zostały załadowane poprawnie.
* Merge data from different sources based on a common key.
* Use a Left Outer Join to add customer details to sales orders.
* Use external help (e.g., ChatGPT) to get guidance on writing custom M code, without copying complete solutions.


*Krok 2: Sprawdź i przekonwertuj typy danych
=== Files Provided ===
Potwierdź, że każda kolumna ma odpowiedni typ danych w obu zapytaniach. Na przykład, zauważ, że kolumna OrderDate w PQ_sales.csv jest zaimportowana jako tekst z powodu niestandardowego formatu. Wskazówka: Użyj funkcji transformacji, jeśli jakiekolwiek dostosowania są potrzebne.
You can download the following two CSV files:


*Krok 3: Scal dane
📂 [[Media:PQ_sales2.csv]] – Contains sales order data 
Scal zapytanie PQ_sales.csv z zapytaniem PQ_customers.csv. Użyj kolumny Customer jako klucza dopasowania. Wybierz lewy zewnętrzny join (Left Outer Join), aby każde zamówienie sprzedaży zostało zachowane wraz z odpowiednimi szczegółami klienta. Propozycja: Jeśli nie wiesz, jak napisać kod M do tego scalenia, zapytaj ChatGPT o wskazówki dotyczące scalania zapytań.


*Krok 4: Przejrzyj scalone dane
📂 [[Media:PQ_customers.csv]] – Contains customer information
Potwierdź, że wynikowe zapytanie zawiera dodatkowe kolumny (np. Region, CustomerSince) z pliku PQ_customers.csv. Sprawdź scalone dane, aby upewnić się, że szczegóły klientów zostały poprawnie połączone z odpowiednimi zamówieniami sprzedaży.


*Krok 5: Zapisz swoją pracę
=== Instructions ===
Zapisz swoje zapytanie i udokumentuj kroki transformacji, które zastosowałeś.


=== Zadanie ===
* Step 1: Import the data 
Open Power Query in Excel or Power BI. 
Import data from both files: PQ_sales.csv and PQ_customers.csv. 
Check that both queries have been loaded correctly.


Wykonaj kroki opisane powyżej w Power Query. Eksperymentuj z interfejsem graficznym i niestandardowym kodem M w celu wykonania scalania. Korzystaj z zasobów zewnętrznych (np. ChatGPT) w celu uzyskania wskazówek lub rozwiązywania problemów, ale unikaj kopiowania kompletnych rozwiązań dosłownie.
* Step 2: Check and convert data types 
Ensure that each column has the correct data type in both queries.
For example, note that the *OrderDate* column in PQ_sales.csv may be imported as text due to a non-standard format.
Tip: Use the transformation functions if any adjustments are needed.


== Moduł 3: Tworzenie niestandardowych kolumn i funkcji ==
* Step 3: Merge the data 
Merge the *PQ_sales.csv* query with the *PQ_customers.csv* query. 
Use the *Customer* column as the matching key. 
Choose the Left Outer Join option so that every sales order is retained along with the corresponding customer details. 
Suggestion: If you’re not sure how to write M code for this merge, ask ChatGPT for tips on how to merge queries.


=== Cel ===
* Step 4: Review the merged data 
W tym ćwiczeniu nauczysz się:
Confirm that the resulting query includes additional columns (e.g., *Region*, *CustomerSince*) from the PQ_customers.csv file. 
Check the merged data to ensure that customer details have been correctly linked to the corresponding sales orders.


Tworzyć niestandardowe kolumny obliczeniowe w Power Query.
* Step 5: Save your work 
Save your query and document the transformation steps you applied.


Używać wbudowanych funkcji Power Query do manipulacji tekstem, liczbami i datami.
=== Task ===


Pisać niestandardowe funkcje w języku M, aby automatyzować transformacje.
Perform the steps described above in Power Query. 
Experiment with both the graphical interface and custom M code to complete the merge. 
Use external resources (e.g., ChatGPT) for guidance or troubleshooting, but avoid copying complete solutions verbatim.


Korzystać z ChatGPT do pomocy przy pisaniu i optymalizacji kodu M.
== Module 3: Creating Custom Columns and Functions ==


=== Dostarczone dane ===
=== Objective ===  
In this exercise, you will learn to:


Do tego ćwiczenia wykorzystamy następujące zbiory danych:
* Create custom calculated columns in Power Query.
* Use built-in Power Query functions to manipulate text, numbers, and dates.
* Write custom M functions to automate transformations.
* Use ChatGPT to assist in writing and optimizing M code.


📂[[Plik:PQ_sales2.csv]] (używany w poprzednich modułach)
=== Files Provided ===


📂[[Plik:PQ_discounts.csv]] (nowy zbiór danych) - zawiera stawki rabatowe w zależności od typu produktu.
The following datasets are used for this exercise:


=== Instrukcje ===
📂[[Media:PQ_sales2.csv]] (used in previous modules) 
📂[[Media:PQ_discounts.csv]] (new dataset) – contains discount rates based on product type.


*Krok 1: Importuj dane
=== Instructions ===
Otwórz Power Query w Excelu lub Power BI.
Zaimportuj oba pliki: PQ_sales.csv i PQ_discounts.csv.
Upewnij się, że obie tabele zostały załadowane poprawnie.


* Step 1: Import the data 
Open Power Query in Excel or Power BI. 
Import both files: *PQ_sales.csv* and *PQ_discounts.csv*. 
Ensure both tables are loaded correctly.


<br>
<br>
*Krok 2: Tworzenie niestandardowej kolumny dla całkowitego kosztu
* Step 2: Create a custom column for total cost 
W tabeli PQ_sales dodaj nową niestandardową kolumnę:
In the *PQ_sales* table, add a new custom column:
Przejdź do Dodaj kolumnę Kolumna niestandardowa.
Go to **Add Column Custom Column**.
Nazwij ją TotalCost.
Name it `TotalCost`.
Utwórz formułę do obliczenia całkowitego kosztu jako:
Create a formula to calculate the total cost as:
  Quantity * Cost
  `Quantity * Cost
Click OK and review the results.


Kliknij OK i sprawdź wyniki.
<br>
* Step 3: Apply discounts using merge 
Merge *PQ_sales* with *PQ_discounts* using the *Product* column as the key.
Expand the `DiscountRate` column into the *PQ_sales* table. 
Add another custom column named `DiscountedPrice`: 
`[TotalCost] - ([TotalCost] * [DiscountRate])` 
Check that the new column correctly applies the discounts.
 
<br>
* Step 4: Create a custom function in M 
Create a function to categorize products into price bands: 
Go to **Home → Advanced Editor**. 
Write an M function that takes `Cost` as input and returns a category: 
Low if Cost < 500  Medium if Cost between 500 and 1500  High if Cost > 1500 




<br>
<br>
*Krok 3: Zastosowanie rabatów za pomocą scalania
* Step 5: Assign categories 
Scal PQ_sales z PQ_discounts, używając kolumny Product jako klucza.
In the *PQ_sales* table, add a custom column using the function. 
Rozwiń kolumnę DiscountRate do tabeli PQ_sales.
Name the column `PriceCategory`.
Dodaj kolejną niestandardową kolumnę o nazwie DiscountedPrice:
Make sure the categories display correctly based on the values in the *Cost* column.
  [TotalCost] - ([TotalCost] * [DiscountRate])
 
=== Task ===
* ✔ Complete all steps in Power Query. 
* ✔ Experiment with both the graphical interface and M code.  
* ✔ Use ChatGPT for troubleshooting or refining your M scripts.


Sprawdź, czy nowa kolumna poprawnie stosuje rabaty.
== Module 4: Advanced Data Transformations in Power Query ==
<br>
*Krok 4: Tworzenie niestandardowej funkcji w M
Utwórz funkcję do kategoryzowania produktów w różne przedziały cenowe:
Przejdź do Strona główna → Edytor zaawansowany.
Napisz funkcję M, która przyjmuje Cost jako wejście i zwraca kategorię:
Low if Cost < 500 Medium if Cost between 500 and 1500 High if Cost > 1500


Wskazówka: Jeśli nie jesteś pewny, jak zbudować funkcję, zapytaj ChatGPT: "Jak napisać funkcję M, która kategoryzuje ceny na Low, Medium i High?" <br>
=== Objective ===
In this module, you will learn:
* ✔ How to pivot and unpivot data in Power Query 
* ✔ How to split and merge columns for better data structure 
* ✔ How to use conditional transformations 
* ✔ How to leverage ChatGPT to build complex M scripts


*Krok 5: Przypisywanie kategorii
=== Files Provided ===
W tabeli PQ_sales dodaj niestandardową kolumnę, używając funkcji.
Nazwij kolumnę PriceCategory.
Upewnij się, że kategorie są wyświetlane poprawnie na podstawie wartości w kolumnie Cost.


=== Zadanie ===
This exercise introduces a new dataset: 
*✔ Wykonaj wszystkie kroki w Power Query.  
'''📂 [[Media:PQ_sales_pivot.csv]]''' – contains monthly sales data in a pivoted format.
*✔ Eksperymentuj zarówno z interfejsem graficznym, jak i z kodem M.
*✔ Korzystaj z ChatGPT w celu rozwiązywania problemów lub poprawiania skryptu M.


== Moduł 4: Zaawansowane transformacje danych w Power Query ==
=== Instructions ===


=== Cel ===
'''Step 1: Import the data''' 
W tym module nauczysz się:  
Open Power Query in Excel or Power BI.
*✔ Jak pivotować i unpivotować dane w Power Query.  
Import the file *PQ_sales_pivot.csv*.
*✔ Jak dzielić i scalać kolumny w celu lepszej strukturyzacji danych.  
Ensure the table loads correctly.
*✔ Jak używać transformacji warunkowych.  
*✔ Jak wykorzystać ChatGPT do tworzenia złożonych skryptów M.


=== Dostarczone dane ===
<br>
'''Step 2: Unpivot the data''' 
The current table has a wide format that is not ideal for analysis. 
Unpivot the monthly columns so the data structure becomes:


Do tego ćwiczenia wprowadzimy nowy zbiór danych:
* Product 
  PQ_sales_pivot.csv, który zawiera miesięczne dane sprzedaży różnych produktów.
* Category 
* Month  
* Sales Amount


'''📂 [[Plik:PQ_sales_pivot.csv]]''' - Struktura danych w formacie pivot.
How to do it:


=== Instrukcje ===
# Click **Transform → Use First Row as Headers** to make sure column names are correct.
# Select the month columns (e.g., Jan 2025, Feb 2025, etc.).
# Click **Transform → Unpivot Columns**.
# Rename the resulting columns: 
## Attribute → Month
## Value → Sales Amount


Krok 1: Importuj dane
Otwórz Power Query w Excelu lub Power BI.
Zaimportuj plik PQ_sales_pivot.csv.
Upewnij się, że tabela została poprawnie załadowana.
<br>
<br>
Krok 2: Unpivotowanie danych
'''Step 3: Splitting and merging columns'''
Obecna tabela ma szeroki format, który nie jest idealny do analizy.
 
Unpivotuj kolumny miesięczne, aby dane miały strukturę:
# The `Month` column now contains values like "Jan 2025".
*Product
# Split this column into `Month Name` and `Year`:
*Category  
# Select the `Month` column. 
*Month
# Click **Transform → Split Column → By Delimiter**. 
*Sales Amount
# Choose space (" ") as the delimiter. 
# Rename the new columns to `Month Name` and `Year`.
 
*Example of merging columns:* 
To merge `Product` and `Category`, select both columns: 
# Click **Transform → Merge Columns**
# Use `" - "` as the separator (e.g., `"Monitor - Electronics"`).
 
<br>
<br>
'''Step 4: Adding conditional transformations''' 
Add a new custom column named `Sales Performance` with the following logic:
if [Sales Amount] < 300 then "Low"
else if [Sales Amount] >= 300 and [Sales Amount] < 800 then "Medium"
else "High"
Make sure the column correctly categorizes the sales performance.
=== Task ===
* ✔ Complete all steps in Power Query 
* ✔ Experiment with unpivoting, splitting, merging, and conditional logic 
* ✔ Use ChatGPT for troubleshooting or refining your M scripts
== Module 5: Parameterization and Dynamic Queries in Power Query ==
=== Objective ===
In this module, you will learn: 
* ✔ How to create parameters in Power Query 
* ✔ How to use parameters for dynamic filtering and query control
=== Files Provided ===
This exercise uses the following files:
* '''📂 [[Media:PQ_sales2.csv]]''' 
* '''📂 [[Media:PQ_parameters.xlsx]]''' – contains values for dynamic filtering
=== Instructions ===
* 🔹 Step 1: Load the CSV file into Power Query 
# Open Excel and go to **Data → Get Data → From File → From Text/CSV** 
# Select the file `PQ_sales.csv` and load it into Power Query 
# Make sure Power Query recognizes the data correctly
* 🔹 Step 2: Process the `Parameters` table 
# In Power Query, go to the **`Parameters`** table 
# **Transpose the table** – click **Transform → Transpose** 
# **Use the first row as headers** – click **Transform → Use First Row as Headers** 
# **Change the data types** for `startDate` and `endDate` to **Date**: 
## Click the `startDate` column header → choose type `Date` 
## Repeat for `endDate`
* 🔹 Step 3: Create separate queries for `startDate` and `endDate` 
# In the `Parameters` table, right-click the value in `startDate` → **Add as New Query** 
# Repeat this for `endDate`
* 🔹 Step 4: Change the data type of `OrderDate` in the `PQ Sales` table to date 
# Go back to the `PQ Sales` query 
# The `OrderDate` column contains dates in `DD MM YY` format 
# **Split the column into three parts**: 
## Click **Transform → Split Column → By Delimiter** 
## Choose **Space** (` `) as the delimiter 
## You will get: `OrderDate.1`, `OrderDate.2`, `OrderDate.3` (day, month, year) 
# **Change their types to `Number` (Int64.Type)** 
# **Merge into proper `YYYY-MM-DD` format**: 
## Click **Merge Columns** 
## Order the columns as: `OrderDate.2`, `OrderDate.3`, `OrderDate.1` (month, year, day) 
## Use `/` as the separator 
## Rename the new column to `DateOrder` 
## Change its type to **Date**
* 🔹 Step 5: Add a dynamic filter to `DateOrder` 
# Open the **Advanced Editor** (`View → Advanced Editor`) 
# Find the last step before `in`, such as:
"Renamed Columns" = Table.RenameColumns(#"Changed Type2",{{"Merged", "DateOrder"}})
* 🔹 Step 6: Add the filter line:
#"Filtered Rows" = Table.SelectRows(#"Renamed Columns", each [DateOrder] >= startDate and [DateOrder] <= endDate)
Ensure that `startDate` and `endDate` are in Date format.
Update the final `in` line to return the filtered table:
in
#"Filtered Rows"
* 🔹 Step 7: Check the results 
# Click **Done** 
# Verify that the data is correctly filtered 
# Click **Close & Load** to load the data into Excel
== Module 6: Automating Data Combining and Refreshing in Power Query ==
=== Objective ===
In this module, you will learn: 
* ✔ How to automatically import and combine files from a folder 
* ✔ How to handle different column names across files 
* ✔ How to prepare data for reporting regardless of source file structure 
* ✔ How to set up automatic data refresh in Power Query
=== Files Provided ===
This exercise uses a set of sales files located in a single folder:
'''📂 [[Media:Sales_Jan.xlsx]]''' – Sales for January 
'''📂 [[Media:Sales_Feb.xlsx]]''' – Sales for February 
'''📂 [[Media:Sales_Mar.xlsx]]''' – Sales for March 
Each file contains similar data, but the sales column names differ:
* In *Sales_Jan.xlsx*: the sales column is named `Total Sale` 
* In *Sales_Feb.xlsx*: the column is named `Revenue` 
* In *Sales_Mar.xlsx*: the column is named `SalesAmount`
The goal is to combine these files into a single dataset and standardize the column names.
=== Instructions ===
* Step 1: Load files from a folder 
# Open Power Query in Excel 
# Go to **Data → Get Data → From File → From Folder** 
# Select the folder containing the files (Sales_Jan.xlsx, Sales_Feb.xlsx, Sales_Mar.xlsx) 
# Click **Load** to add files to Power Query without combining them automatically
* Step 2: Use M code to load the data 
# Open **Advanced Editor** in Power Query 
# Paste the following M code and click **Done**:
let
// Load files from folder
Source = Folder.Files("C:\Users\pathToFolder..."),
// Add a column to access the Excel file contents
AddContent = Table.AddColumn(Source, "Custom", each Excel.Workbook([Content])),
/ Expand content to view all data
ExpandContent = Table.ExpandTableColumn(AddContent, "Custom", {"Name", "Data"}, {"File Name", "Data"})
in
ExpandContent
# After applying the code, you will see a new `Data` column
* Step 3: Expand the table contents 
# Click the expand icon next to the `Data` column 
# This reveals the full data from each file 
# Ensure that all relevant columns from all files are visible
* Step 4: Remove unnecessary columns 
# Review the table and remove technical columns (e.g., file path) not needed for analysis 
# Go to **Transform → Remove Columns** and select what to discard
* Step 5: Rename columns 
# Rename the varying sales columns to a consistent name (e.g., `Sales`) 
# Use **Transform → Rename Column** to apply a uniform structure
* Step 6: Remove unnecessary rows (e.g., repeated headers) 
# Apply a filter on the column containing sales values 
# Remove rows with repeated headers caused by merging files 
# Go to **Transform → Remove Rows → Remove Duplicates**, or filter manually
* Step 7: Enable automatic refresh 
# Go to **Data → Query Properties → Refresh data when opening the file** 
# Optionally set automatic refresh every X minutes 
# If a new file (e.g., *Sales_Apr.xlsx*) is added to the folder, Power Query will automatically include it upon refresh!
=== Task ===
* ✔ Load and combine data from *Sales_Jan.xlsx*, *Sales_Feb.xlsx*, and *Sales_Mar.xlsx* 
* ✔ Standardize column names and format the data consistently 
* ✔ Remove empty rows, unnecessary columns, and duplicates 
* ✔ Set up auto-refresh so new files are included automatically 
* ✔ Use ChatGPT to optimize the M code in Power Query
== Module 7: Optimizing Query Performance in Power Query ==
=== Objective ===
In this module, you will learn: 
* ✔ How to speed up Power Query when working with large datasets 
* ✔ How to avoid inefficient operations that slow down queries 
* ✔ How to use buffering and database-level transformations 
* ✔ How to minimize the amount of data processed for better performance
=== Introduction ===
Power Query enables powerful data transformation, but with large datasets, performance can suffer. In this module, you will learn best practices to reduce query execution time.
=== Instructions ===
* Step 1: Avoid unnecessary operations on the entire dataset 
# Load a large CSV file: 📂[[Media:PQSales_Large.csv]] 
# Check the number of rows and columns – the more data, the more important the optimization 
# Remove unnecessary columns at the beginning of the query instead of the end 
# Apply early filtering to reduce the number of rows right after import
* Step 2: Use buffering (Table.Buffer) 
# Understand how step-by-step processing works – each operation may cause Power Query to recalculate previous steps 
# Add `Table.Buffer()` after the filter step to avoid re-processing:
let
Source = Csv.Document(File.Contents("C:\Users\gp\Desktop\PQ\Sales_Large.csv"),[Delimiter=",", Columns=6, Encoding=1252, QuoteStyle=QuoteStyle.None]),
FilteredRows = Table.SelectRows(Source, each [Cost] > 500),
BufferedData = Table.Buffer(FilteredRows)
in
BufferedData
Using `Table.Buffer()` ensures that the results are stored in memory and not recalculated at each step.
* Step 3: Minimize the number of loaded rows 
# When working with large databases or CSV files, load only the needed columns and rows 
# Use **Keep Top Rows** to load e.g., the first 1000 rows for testing 
# Apply **Remove Duplicates** early to reduce the volume of data being processed
* Step 4: Optimize database connections 
# If working with SQL Server, Power BI, or another database, avoid importing full tables into Power Query 
# Instead, apply filtering and grouping on the database side using native SQL
Example:
let
Source = Sql.Database("ServerName", "DatabaseName", [Query="SELECT OrderID, OrderDate, Customer, Product FROM Sales WHERE Cost > 500"])
in
Source
This ensures Power Query pulls only the filtered data instead of processing the entire table in memory.
* Step 5: Avoid "drill-down" operations on large datasets 
# Power Query often suggests drill-downs (e.g., selecting a single value from a table) 
# When working with large data, operate on whole tables instead of individual records
* Step 6: Automatically refresh optimized queries 
# Once optimized, configure the query to refresh regularly 
# In Excel, go to **Data → Query Properties → Refresh data when opening the file**
=== Task ===
* Load the large CSV file (*PQSales_Large.csv*) 
* Limit the number of loaded rows and columns 
* Apply `Table.Buffer()` and observe performance improvements 
* If using a database, optimize your SQL query 
* Set up auto-refresh for the optimized query 
* Use ChatGPT to analyze performance and further optimize M code
== Module 8: Creating Dynamic Reports and Dashboards in Excel with Power Query ==
=== Objective === 
In this module, you will learn: 
* ✔ How to use Power Query to dynamically generate reports 
* ✔ How to combine data from multiple sources into a single report 
* ✔ How to create interactive reports using PivotTables 
* ✔ How to automate report refreshing in Excel
=== Files Provided === 
The following files are used for this exercise: 
* 📂 [[Media:PQ_Sales_Data.xlsx]] – Sales data 
* 📂 [[Media:PQ_Regions.xlsx]] – Sales regions 
* 📂 [[Media:PQ_Targets.xlsx]] – Sales targets
=== Instructions ===
'''Step 1: Import and combine data sources'''
# Open Power Query in Excel 
# Import the files: *PQ_Sales_Data.xlsx*, *PQ_Regions.xlsx*, and *PQ_Targets.xlsx* 
# Merge the data using a common key – for example, the `Region` column 
# Verify that the data is correctly combined and properly formatted
'''Step 2: Create a dynamic report'''
# Click **Close & Load To...** and select **Pivot Table** 
# Insert the PivotTable in a new worksheet, using the Power Query output as the source 
# In the PivotTable Fields pane, set:
  * Rows → `Region` 
  * Columns → `Month` 
  * Values → `Sum of Sales`
Check for accuracy and apply formatting to the table
'''Step 3: Add conditional formatting'''
# Select the `Sum of Sales` column in the PivotTable 
# Go to **Conditional Formatting → Color Scales** 
# Apply gradient colors to highlight low and high sales values 
# Add a rule: “Greater than” and highlight values above the sales target (from *PQ_Targets.xlsx*) using:
=B5 > VLOOKUP($A5,Targets!$A$2:$B$8,2,0)


Jak to zrobić:
'''Step 4: Automate data refreshing'''


#Kliknij Transformuj Użyj pierwszego wiersza jako nagłówków, aby upewnić się, że kolumny mają poprawne nazwy.
# Go to **Data Query Properties → Refresh data when opening the file** 
#Zaznacz kolumny z miesiącami (Jan 2025, Feb 2025 itd.).
# Optionally set auto-refresh every X minutes 
#Kliknij Transformuj → Unpivotuj kolumny.
# Test the report by updating the source files and verifying that the report refreshes correctly
#Zmień nazwy wynikowych kolumn:
 
"Attribute" → Month
=== Task ===
"Value" → Sales Amount
 
<br>
* ✔ Load and combine data from *PQ_Sales_Data.xlsx*, *PQ_Regions.xlsx*, and *PQ_Targets.xlsx* 
Krok 3: Dzielnie i scalanie kolumn
* ✔ Create a PivotTable and format it dynamically 
#Kolumna Month obecnie zawiera wartości takie jak "Jan 2025".
* ✔ Add conditional formatting based on sales targets 
#Podziel tę kolumnę na Month Name i Year:
* ✔ Set up automatic data refreshing 
#Zaznacz kolumnę Month.
* ✔ Use ChatGPT to analyze and optimize Power Query transformations
#Kliknij Transformuj → Podziel kolumnę → Według ogranicznika.
 
#Wybierz spację (" ") jako ogranicznik.
== Summary: Modules 1–8 ==
#Zmień nazwy nowych kolumn na Month Name i Year.
 
#Przykład scalania kolumn:
=== Objective ===
In this exercise, you will summarize all the key concepts learned so far in Power Query by performing a series of transformations on inventory and supplier data. 
You will apply data import, filtering, merging, column creation, custom functions, and query optimization.
 
=== Files Provided ===
The following files are used for this exercise: 
* '''📂 [[Media:PQ_inventory.csv]]''' – Inventory stock data 
* '''📂 [[Media:PQ_suppliers.csv]]''' – Supplier information 
* '''📂 [[Media:PQ_orders.csv]]''' – Warehouse delivery orders
 
=== Instructions ===
 
'''🔹 Step 1: Import data''' 
# Open Power Query in Excel or Power BI 
# Import the three CSV files: `PQ_inventory.csv`, `PQ_suppliers.csv`, `PQ_orders.csv` 
# Make sure all datasets are loaded correctly
 
'''🔹 Step 2: Check and convert data types''' 
# Ensure all columns in each dataset have correct data types 
# Issue to solve: the `StockLevel` column was incorrectly imported as text because it contains units like `150 kg`, `200 l`, `75 pcs` 
# Transform the `StockLevel` column to extract numeric values and store the unit in a new column `Unit` 
# Verify that the `SupplierID` column is recognized as an integer
 
'''🔹 Step 3: Merge data''' 
# Merge `PQ_inventory.csv` with `PQ_suppliers.csv` using the `SupplierID` key 
# Use a **Left Outer Join** to retain all inventory records 
# Then merge `PQ_orders.csv` with `PQ_inventory.csv` using the `ProductID` key 
# Verify that supplier and order info have been successfully added to the inventory table
 
'''🔹 Step 4: Create custom columns''' 
# Add a column `ReorderLevel` that flags products needing restocking when `StockLevel` is less than `MinimumStock` 
# Add a column `DaysSinceLastOrder` that calculates the number of days since the last order for each product 
# Create a custom M function that assigns order priority:
 
<pre>
if [StockLevel] < [MinimumStock] and [DaysSinceLastOrder] > 30 then "High"
else if [StockLevel] < [MinimumStock] then "Medium"
else "Low"
</pre>


Jeśli chcesz scalić Product i Category, zaznacz obie kolumny.
# Add a column `OrderPriority` and apply this function


#Kliknij Transformuj → Scal kolumny.
'''🔹 Step 5: Conditional filtering and transformations''' 
#Użyj " - " jako separatora (np. "Monitor - Electronics").
# Remove products that have a `Discontinued` status 
<br>
# Add a new column `SupplierRating` that classifies suppliers by reliability:
Krok 4: Dodanie transformacji warunkowych
*Dodaj nową niestandardową kolumnę o nazwie "Sales Performance":


if [Sales Amount] < 300 then "Low" else if [Sales Amount] >= 300 and [Sales Amount] < 800 then "Medium" else "High"
<pre>
if [OnTimeDeliveryRate] < 80 then "Excellent"  
else if [OnTimeDeliveryRate] <= 90 then "Good"  
else "Poor"
</pre>


*Upewnij się, że kolumna poprawnie kategoryzuje wyniki sprzedaży.
# Verify that the rating logic works correctly


=== Zadanie ===
'''🔹 Step 6: Reshape the data structure''' 
*✔ Wykonaj wszystkie kroki w Power Query.
# Unpivot columns `Stock_Jan`, `Stock_Feb`, `Stock_Mar` into: `Product`, `Month`, `Stock Level` 
*✔ Eksperymentuj z unpivotowaniem, dzieleniem, scalaniem i logiką warunkową.
# Split the `ProductDetails` column into `ProductName` and `Category` 
*✔ Korzystaj z ChatGPT w celu rozwiązywania problemów lub poprawiania skryptu M.
# Merge the `SupplierName` and `Country` columns using `" - "` as a separator


== Moduł 5: Parametryzacja i dynamiczne zapytania w Power Query ==
'''🔹 Step 7: Query optimization''' 
# Apply `Table.Buffer()` to improve performance 
# Remove unused columns and duplicates at the beginning of the transformations, not at the end 
# If working with large data, limit the loaded rows to a test sample of 1000


=== Cel ćwiczenia ===
'''🔹 Step 8: Export results''' 
W tym module nauczysz się:
# Load the final query as a table into Excel 
*✔ Jak tworzyć parametry w Power Query.
# Test data refresh by updating source files 
*✔ Jak używać parametrów do dynamicznego filtrowania i sterowania wynikami zapytań.
# Set up auto-refresh for the query
*✔ Jak budować dynamiczne źródła danych w oparciu o dane wejściowe użytkownika.*✔ Jak wykorzystać ChatGPT do pisania i optymalizacji kodu M dla parametrów.


=== Dostarczone dane ===
=== Task ===
W tym ćwiczeniu użyjemy pliku PQ_sales.csv oraz nowego zestawu danych PQ_parameters.xlsx, który zawiera wartości do dynamicznego filtrowania.
* ✔ Complete all the steps listed above 
* ✔ Experiment with both the graphical interface and M code 
* ✔ Apply query optimization to improve performance 
* ✔ Ensure all transformations are correct and results are as expected 
* ✔ Use ChatGPT for troubleshooting or optimizing your M script


'''📂 [[File:PQ_parameters.xlsx]]''' - Tabela z parametrami do filtrowania:
== Module 9: Importing and Analyzing PDF Files in Power Query ==


=== Instrukcje ===
=== Objective ===
In this module, you will learn:


#Krok 1: Importowanie danych
* ✔ How to import data from PDF files into Power Query
#Otwórz Power Query w Excelu lub Power BI.
* ✔ How to transform data and perform analysis on business reports 
#Zaimportuj oba zestawy danych:
* ✔ How to visualize results and draw insights from reports
PQ_sales.csv (Dane transakcji sprzedaży).
PQ_parameters.xlsx (Parametry filtrowania).
#Upewnij się, że tabele zostały poprawnie załadowane.
*Krok 2: Tworzenie i używanie parametrów dynamicznie
#Załaduj tabelę PQ_parameters.xlsx do Power Query.
#Upewnij się, że tabela zawiera dwie kolumny: ParameterName (Nazwa parametru) i Value (Wartość).
#Kliknij Przekształć → Użyj pierwszego wiersza jako nagłówków, aby poprawnie przypisać nazwy kolumn.
#Przekształć tabelę w rekord, aby łatwo odwoływać się do wartości:
#Zaznacz kolumnę ParameterName i wykonaj Przestawienie kolumny.
#Kliknij Przekształć → Przestaw kolumnę, ustawiając Value jako kolumnę wartości.
Powinno to utworzyć rekord zawierający pola StartDate, EndDate, MinCost, MaxCost, FilePath (jeśli chcemy dynamicznie pobierać dane z innego pliku).
#Upewnij się, że wszystkie wartości mają poprawny typ danych (Data dla StartDate i EndDate, Liczba dla MinCost i MaxCost, Tekst dla FilePath).
Teraz parametry są dynamicznie pobierane z pliku PQ_parameters.xlsx.
*Krok 3: Zastosowanie parametrów do filtrowania danych
Filtrowanie danych sprzedaży według zakresu dat:
#Otwórz zapytanie PQ_sales w Power Query.
#Upewnij się, że kolumna OrderDate ma typ Data. Jeśli nie, zmień go na Data poprzez Przekształć → Typ danych → Data.
#Kliknij ikonę filtra w kolumnie OrderDate.
#Wybierz Filtry dat → Filtr niestandardowy.
#W oknie filtrów ustaw:
#Pierwszy warunek na jest po lub równe i odwołaj się do StartDate z rekordu.
#Drugi warunek na jest przed lub równe i odwołaj się do EndDate z rekordu.
#Kliknij OK, aby zastosować filtr.
Alternatywnie, możesz użyć wyrażenia M do filtrowania danych:


Table.SelectRows(PQ_sales, each [OrderDate] >= Parameters[StartDate] and [OrderDate] <= Parameters[EndDate])
=== Files Provided ===
The following PDF reports are used in this exercise:


Filtrowanie danych sprzedaży według zakresu kosztów:
* 📂 [[Media:Monthly_Sales_Report_Jan2024.pdf]] – Sales Report 
* 📂 [[Media:Employee_Attendance_Q1_2024.pdf]] – Employee Attendance Report 
* 📂 [[Media:Customer_Feedback_Survey_2024.pdf]] – Customer Feedback Report


#Upewnij się, że kolumna Cost ma typ Liczba.
=== Instructions ===
#Kliknij ikonę filtra w kolumnie Cost.
#Wybierz Filtry liczb → Pomiędzy.
#Ustaw pierwszy warunek na większe niż lub równe i odwołaj się do MinCost z rekordu.
#Ustaw drugi warunek na mniejsze niż lub równe i odwołaj się do MaxCost z rekordu.
#Kliknij OK, aby zastosować filtr.


Alternatywnie, możesz użyć wyrażenia M:
🔹 '''Task 1: Sales Report Analysis'''


  Table.SelectRows(PQ_sales, each [Cost] >= Parameters[MinCost] and [Cost] <= Parameters[MaxCost])
# Calculate total sales for all products  
# Identify the product with the highest and lowest sales 
# Compute the average transaction value based on transaction count and total sales 
# Group data by region and calculate total sales per region 
# Create a pivot table showing sales by region and product


*Krok 4: Tworzenie dynamicznego źródła danych
🔹 '''Task 2: Employee Attendance Report Analysis'''


Dynamiczne pobieranie danych z pliku o ścieżce przechowywanej w parametrze:
# Calculate the average attendance rate across all departments 
# Identify the department with the highest and lowest attendance 
# Add a new column classifying attendance into categories:
## High: above 95% 
## Medium: 85%–95% 
## Low: below 85% 
# Filter the data to show only employees with low attendance 
# Create a bar chart showing average attendance by department


Source = Excel.Workbook(File.Contents(Parameters[FilePath]), null, true)
🔹 '''Task 3: Customer Feedback Report Analysis'''


Jeśli chcemy pobierać dane z pliku, którego ścieżka może się zmieniać, możemy dodać parametr FilePath do PQ_parameters.xlsx.
# Calculate the average customer rating on a 1–5 scale 
# Count how many customers gave a rating of 1 or 5 
# Generate a summary report of the most frequent positive and negative comments 
# Sort the data by customer rating from lowest to highest 
# Create a pie chart showing the distribution of customer ratings


FilePath powinien zawierać pełną ścieżkę do pliku, np. C:\Dane\Raport.xlsx.
=== Summary ===
* ✔ Complete the analysis tasks for each report separately 
* ✔ Apply filtering, sorting, and grouping operations 
* ✔ Use pivot tables to aggregate data 
* ✔ Visualize results using charts in Excel or Power BI 
* ✔ Use ChatGPT if you encounter difficulties during analysis


Po aktualizacji wartości w PQ_parameters.xlsx, Power Query automatycznie załaduje dane z nowej lokalizacji.
== Module 10: Importing and Analyzing Web Data in Power Query ==


=== Zadanie ===
=== Objective ===
*✔ Stwórz i wykorzystaj parametry do dynamicznego filtrowania danych.
In this module, you will learn:
*✔ Zaimplementuj dynamiczne źródło danych wykorzystujące parametryzowaną ścieżkę pliku.
*✔ Skorzystaj z ChatGPT do analizy i optymalizacji kodu M w Power Query.


== Moduł 6: Automatyzacja Przekształceń i Odświeżania Danych w Power Query ==
* ✔ How to import data from statistical tables available on Wikipedia into Power Query
* ✔ How to transform and analyze data about countries of the world 
* ✔ How to visualize comparison results in Excel or Power BI


=== Cel ćwiczenia ===
=== Data Sources ===
In this exercise, we’ll use real tabular data about countries of the world imported directly from Wikipedia. These include:


W tym module nauczysz się:
* 📊 **Surface area of countries** 🌍 
*✔ Jak tworzyć kroki przekształceń, które automatycznie aktualizują się po dodaniu nowych danych.
* 📊 **Population by country** 👥 
*✔ Jak korzystać z funkcji Table.Buffer() i innych metod optymalizacji wydajności.
* 📊 **Gross Domestic Product (GDP) by country** 💰 
*✔ Jak ustawić automatyczne odświeżanie danych w Power Query.
*✔ Jak wykorzystać ChatGPT do analizy i optymalizacji kodu M.


=== Dostarczone dane ===
'''Sources:''' 
* [https://en.wikipedia.org/wiki/List_of_countries_and_dependencies_by_area List of countries by area] 
* [https://en.wikipedia.org/wiki/List_of_countries_and_dependencies_by_population List of countries by population] 
* [https://en.wikipedia.org/wiki/List_of_countries_by_GDP_(nominal) List of countries by nominal GDP]


W tym ćwiczeniu użyjemy pliku PQ_sales_dynamic.xlsx, który zawiera regularnie aktualizowane dane sprzedażowe.
=== Instructions ===


'''📂 [[File:PQ_sales_dynamic.xlsx]]''' – Dynamicznie aktualizowane dane sprzedaży
🔹 '''Step 1: Import data from Wikipedia'''


=== Instrukcje ===
# Open Power Query in Excel or Power BI 
# Choose **Get Data → From Web** 
# Enter the URL of one of the Wikipedia pages above 
# Once the available tables are loaded, select the one containing statistical data (e.g., country area, population, or GDP) 
# Click **Load to Power Query** to begin transforming the data


*Krok 1: Importowanie dynamicznych danych
🔹 '''Step 2: Transform and clean the data'''


#Otwórz Power Query w Excelu.
# **Remove unnecessary columns**, keeping only those relevant for analysis 
#Zaimportuj plik PQ_sales_dynamic.xlsx.
# **Change data types** so that numbers are correctly interpreted (e.g., `Area` as number, `GDP` as currency) 
#Upewnij się, że tabela załadowała się poprawnie.
# **Remove empty values** and correct any errors 
#Sprawdź, czy nazwy kolumn w pliku mogą się zmieniać w przyszłości (np. nowe miesiące dodawane do arkusza).
# **Rename columns** to clearer names, such as `Country`, `Area (km²)`, `Population`, `GDP (billion USD)`


*Krok 2: Automatyczne wykrywanie i dynamiczne przekształcenia
🔹 '''Step 3: Analyze and compare countries'''


#Jeśli dane zawierają zmieniające się kolumny, zastosuj Przestawienie kolumn, aby dane były w układzie długim:
# **Calculate population density** by adding a new column with the formula:
#Wybierz wszystkie kolumny oprócz OrderID, Customer, Product.
Population Density = Population / Area 
#Kliknij Przekształć → Przestaw kolumny (Unpivot Columns).
#Jeśli plik zawiera puste wiersze lub kolumny, usuń je:
#Kliknij Przekształć → Usuń wiersze → Usuń puste wiersze.
#Kliknij Przekształć → Usuń kolumny → Usuń puste kolumny.


Upewnij się, że typy danych zostały poprawnie wykryte i dostosowane.
# **Sort countries by GDP** to identify the richest and poorest nations 
# **Compare area and population** to find the largest, smallest, and most populated countries 
# **Apply filtering** to display only selected continents or world regions


*Krok 3: Optymalizacja zapytań i wykorzystanie Table.Buffer()
🔹 '''Step 4: Visualize the results'''
#Zapobiegaj wielokrotnemu ponownemu obliczaniu kroków przez użycie Table.Buffer():
BufferedTable = Table.Buffer(PreviousStep)
#Stosuj tę funkcję, gdy Power Query wykonuje kilka operacji na tej samej tabeli.
#Redukuj liczbę wierszy przed skomplikowanymi operacjami:
#Zastosuj filtrowanie jak najwcześniej, aby ograniczyć liczbę przetwarzanych wierszy.
*Krok 4: Automatyczne odświeżanie danych


#Przejdź do Dane → Właściwości zapytania → Odświeżaj dane przy otwieraniu pliku.
# Create a **pivot table** in Excel to compare area, population, and GDP 
#Możesz także ustawić automatyczne odświeżanie co X minut.
# Insert a **bar chart** to show the largest economies 
# Use a **heat map** to illustrate population density by region 
# Add **conditional formatting** to highlight countries with extreme statistical values


=== Zadanie ===
=== Task ===
*✔ Załaduj i przekształć dynamiczne dane z pliku PQ_sales_dynamic.xlsx.
* ✔ Import world country data from Wikipedia into Power Query 
*✔ Zastosuj dynamiczne przestawienie kolumn, jeśli dane mogą się zmieniać.
* ✔ Transform and clean the data to make it analysis-ready 
*✔ Wykorzystaj Table.Buffer() dla optymalizacji wydajności.
* ✔ Calculate population density and other statistical indicators 
*✔ Ustaw automatyczne odświeżanie danych w Excelu.
* ✔ Create charts and comparison tables in Excel or Power BI 
*✔ Skorzystaj z ChatGPT do optymalizacji kodu M w Power Query.
* ✔ Use ChatGPT if you encounter issues during import or analysis

Latest revision as of 12:47, 24 June 2025

Module 1: Introduction to Power Query and Basic Transformations

Objective

In this exercise, you will learn to:

  • Import data from a CSV file into Power Query.
  • Review and adjust data types, focusing on converting a date stored as text into a proper date type.
  • Apply basic data filtering.
  • Use external help (e.g., ChatGPT) to get guidance on creating custom M code, without directly copying solutions.

Files Provided

You can download the CSV file 📂Media:PQsales.csv which contains sales order data.

Instructions

  • Step 1: Import the data
  1. Open Power Query in Excel or Power BI.
  2. Import the data from the file sales.csv.
  3. Note that the *OrderDate* column was imported as text due to its format (dd/MM/yyyy).
  • Step 2: Check and convert data types
  1. Verify that each column has the correct data type.
  2. Manually transform the *OrderDate* column from text to date type.
  3. Tip: If you encounter difficulties, you can ask ChatGPT for hints on how to write an M function to convert text to date.
  • Step 3: Apply basic filtering
  1. Filter the dataset to keep only rows where the cost is greater than 200.
  2. Suggestion: Use Power Query’s graphical interface or write a simple M script to apply the filter. If needed, consult ChatGPT for implementation ideas.
  • Step 4: Review and save your work
  1. Confirm that the transformations were applied correctly by reviewing the data preview.
  2. Save the query and document the steps you have taken.

Task

Perform the steps described above in Power Query. Experiment with available transformation options and try to understand how each step affects your data. Use ChatGPT for hints or troubleshooting, but avoid copying complete solutions verbatim.

Module 2: Combining and Merging Data from Multiple Sources

Objective

In this exercise, you will learn to:

  • Import data from multiple CSV files into Power Query.
  • Merge data from different sources based on a common key.
  • Use a Left Outer Join to add customer details to sales orders.
  • Use external help (e.g., ChatGPT) to get guidance on writing custom M code, without copying complete solutions.

Files Provided

You can download the following two CSV files:

📂 Media:PQ_sales2.csv – Contains sales order data

📂 Media:PQ_customers.csv – Contains customer information

Instructions

  • Step 1: Import the data

Open Power Query in Excel or Power BI. Import data from both files: PQ_sales.csv and PQ_customers.csv. Check that both queries have been loaded correctly.

  • Step 2: Check and convert data types

Ensure that each column has the correct data type in both queries. For example, note that the *OrderDate* column in PQ_sales.csv may be imported as text due to a non-standard format. Tip: Use the transformation functions if any adjustments are needed.

  • Step 3: Merge the data

Merge the *PQ_sales.csv* query with the *PQ_customers.csv* query. Use the *Customer* column as the matching key. Choose the Left Outer Join option so that every sales order is retained along with the corresponding customer details. Suggestion: If you’re not sure how to write M code for this merge, ask ChatGPT for tips on how to merge queries.

  • Step 4: Review the merged data

Confirm that the resulting query includes additional columns (e.g., *Region*, *CustomerSince*) from the PQ_customers.csv file. Check the merged data to ensure that customer details have been correctly linked to the corresponding sales orders.

  • Step 5: Save your work

Save your query and document the transformation steps you applied.

Task

Perform the steps described above in Power Query. Experiment with both the graphical interface and custom M code to complete the merge. Use external resources (e.g., ChatGPT) for guidance or troubleshooting, but avoid copying complete solutions verbatim.

Module 3: Creating Custom Columns and Functions

Objective

In this exercise, you will learn to:

  • Create custom calculated columns in Power Query.
  • Use built-in Power Query functions to manipulate text, numbers, and dates.
  • Write custom M functions to automate transformations.
  • Use ChatGPT to assist in writing and optimizing M code.

Files Provided

The following datasets are used for this exercise:

📂Media:PQ_sales2.csv (used in previous modules) 📂Media:PQ_discounts.csv (new dataset) – contains discount rates based on product type.

Instructions

  • Step 1: Import the data

Open Power Query in Excel or Power BI. Import both files: *PQ_sales.csv* and *PQ_discounts.csv*. Ensure both tables are loaded correctly.


  • Step 2: Create a custom column for total cost

In the *PQ_sales* table, add a new custom column: Go to **Add Column → Custom Column**. Name it `TotalCost`. Create a formula to calculate the total cost as:

`Quantity * Cost`  

Click OK and review the results.


  • Step 3: Apply discounts using merge

Merge *PQ_sales* with *PQ_discounts* using the *Product* column as the key. Expand the `DiscountRate` column into the *PQ_sales* table. Add another custom column named `DiscountedPrice`:

`[TotalCost] - ([TotalCost] * [DiscountRate])`  

Check that the new column correctly applies the discounts.


  • Step 4: Create a custom function in M

Create a function to categorize products into price bands: Go to **Home → Advanced Editor**. Write an M function that takes `Cost` as input and returns a category:

Low if Cost < 500   Medium if Cost between 500 and 1500  High if Cost > 1500  



  • Step 5: Assign categories

In the *PQ_sales* table, add a custom column using the function. Name the column `PriceCategory`. Make sure the categories display correctly based on the values in the *Cost* column.

Task

  • ✔ Complete all steps in Power Query.
  • ✔ Experiment with both the graphical interface and M code.
  • ✔ Use ChatGPT for troubleshooting or refining your M scripts.

Module 4: Advanced Data Transformations in Power Query

Objective

In this module, you will learn:

  • ✔ How to pivot and unpivot data in Power Query
  • ✔ How to split and merge columns for better data structure
  • ✔ How to use conditional transformations
  • ✔ How to leverage ChatGPT to build complex M scripts

Files Provided

This exercise introduces a new dataset: 📂 Media:PQ_sales_pivot.csv – contains monthly sales data in a pivoted format.

Instructions

Step 1: Import the data Open Power Query in Excel or Power BI. Import the file *PQ_sales_pivot.csv*. Ensure the table loads correctly.


Step 2: Unpivot the data The current table has a wide format that is not ideal for analysis. Unpivot the monthly columns so the data structure becomes:

  • Product
  • Category
  • Month
  • Sales Amount

How to do it:

  1. Click **Transform → Use First Row as Headers** to make sure column names are correct.
  2. Select the month columns (e.g., Jan 2025, Feb 2025, etc.).
  3. Click **Transform → Unpivot Columns**.
  4. Rename the resulting columns:
    1. Attribute → Month
    2. Value → Sales Amount


Step 3: Splitting and merging columns

  1. The `Month` column now contains values like "Jan 2025".
  2. Split this column into `Month Name` and `Year`:
  3. Select the `Month` column.
  4. Click **Transform → Split Column → By Delimiter**.
  5. Choose space (" ") as the delimiter.
  6. Rename the new columns to `Month Name` and `Year`.
  • Example of merging columns:*

To merge `Product` and `Category`, select both columns:

  1. Click **Transform → Merge Columns**.
  2. Use `" - "` as the separator (e.g., `"Monitor - Electronics"`).


Step 4: Adding conditional transformations Add a new custom column named `Sales Performance` with the following logic:

if [Sales Amount] < 300 then "Low"
else if [Sales Amount] >= 300 and [Sales Amount] < 800 then "Medium"
else "High"


Make sure the column correctly categorizes the sales performance.

Task

  • ✔ Complete all steps in Power Query
  • ✔ Experiment with unpivoting, splitting, merging, and conditional logic
  • ✔ Use ChatGPT for troubleshooting or refining your M scripts

Module 5: Parameterization and Dynamic Queries in Power Query

Objective

In this module, you will learn:

  • ✔ How to create parameters in Power Query
  • ✔ How to use parameters for dynamic filtering and query control

Files Provided

This exercise uses the following files:

Instructions

  • 🔹 Step 1: Load the CSV file into Power Query
  1. Open Excel and go to **Data → Get Data → From File → From Text/CSV**
  2. Select the file `PQ_sales.csv` and load it into Power Query
  3. Make sure Power Query recognizes the data correctly
  • 🔹 Step 2: Process the `Parameters` table
  1. In Power Query, go to the **`Parameters`** table
  2. **Transpose the table** – click **Transform → Transpose**
  3. **Use the first row as headers** – click **Transform → Use First Row as Headers**
  4. **Change the data types** for `startDate` and `endDate` to **Date**:
    1. Click the `startDate` column header → choose type `Date`
    2. Repeat for `endDate`
  • 🔹 Step 3: Create separate queries for `startDate` and `endDate`
  1. In the `Parameters` table, right-click the value in `startDate` → **Add as New Query**
  2. Repeat this for `endDate`
  • 🔹 Step 4: Change the data type of `OrderDate` in the `PQ Sales` table to date
  1. Go back to the `PQ Sales` query
  2. The `OrderDate` column contains dates in `DD MM YY` format
  3. **Split the column into three parts**:
    1. Click **Transform → Split Column → By Delimiter**
    2. Choose **Space** (` `) as the delimiter
    3. You will get: `OrderDate.1`, `OrderDate.2`, `OrderDate.3` (day, month, year)
  4. **Change their types to `Number` (Int64.Type)**
  5. **Merge into proper `YYYY-MM-DD` format**:
    1. Click **Merge Columns**
    2. Order the columns as: `OrderDate.2`, `OrderDate.3`, `OrderDate.1` (month, year, day)
    3. Use `/` as the separator
    4. Rename the new column to `DateOrder`
    5. Change its type to **Date**
  • 🔹 Step 5: Add a dynamic filter to `DateOrder`
  1. Open the **Advanced Editor** (`View → Advanced Editor`)
  2. Find the last step before `in`, such as:
"Renamed Columns" = Table.RenameColumns(#"Changed Type2",Template:"Merged", "DateOrder")
  • 🔹 Step 6: Add the filter line:
#"Filtered Rows" = Table.SelectRows(#"Renamed Columns", each [DateOrder] >= startDate and [DateOrder] <= endDate)


Ensure that `startDate` and `endDate` are in Date format.

Update the final `in` line to return the filtered table:

in
#"Filtered Rows"


  • 🔹 Step 7: Check the results
  1. Click **Done**
  2. Verify that the data is correctly filtered
  3. Click **Close & Load** to load the data into Excel

Module 6: Automating Data Combining and Refreshing in Power Query

Objective

In this module, you will learn:

  • ✔ How to automatically import and combine files from a folder
  • ✔ How to handle different column names across files
  • ✔ How to prepare data for reporting regardless of source file structure
  • ✔ How to set up automatic data refresh in Power Query

Files Provided

This exercise uses a set of sales files located in a single folder:

📂 Media:Sales_Jan.xlsx – Sales for January 📂 Media:Sales_Feb.xlsx – Sales for February 📂 Media:Sales_Mar.xlsx – Sales for March

Each file contains similar data, but the sales column names differ:

  • In *Sales_Jan.xlsx*: the sales column is named `Total Sale`
  • In *Sales_Feb.xlsx*: the column is named `Revenue`
  • In *Sales_Mar.xlsx*: the column is named `SalesAmount`

The goal is to combine these files into a single dataset and standardize the column names.

Instructions

  • Step 1: Load files from a folder
  1. Open Power Query in Excel
  2. Go to **Data → Get Data → From File → From Folder**
  3. Select the folder containing the files (Sales_Jan.xlsx, Sales_Feb.xlsx, Sales_Mar.xlsx)
  4. Click **Load** to add files to Power Query without combining them automatically
  • Step 2: Use M code to load the data
  1. Open **Advanced Editor** in Power Query
  2. Paste the following M code and click **Done**:
let
// Load files from folder
Source = Folder.Files("C:\Users\pathToFolder..."),

// Add a column to access the Excel file contents
AddContent = Table.AddColumn(Source, "Custom", each Excel.Workbook([Content])),

/ Expand content to view all data
ExpandContent = Table.ExpandTableColumn(AddContent, "Custom", {"Name", "Data"}, {"File Name", "Data"})

in
ExpandContent


  1. After applying the code, you will see a new `Data` column
  • Step 3: Expand the table contents
  1. Click the expand icon next to the `Data` column
  2. This reveals the full data from each file
  3. Ensure that all relevant columns from all files are visible
  • Step 4: Remove unnecessary columns
  1. Review the table and remove technical columns (e.g., file path) not needed for analysis
  2. Go to **Transform → Remove Columns** and select what to discard
  • Step 5: Rename columns
  1. Rename the varying sales columns to a consistent name (e.g., `Sales`)
  2. Use **Transform → Rename Column** to apply a uniform structure
  • Step 6: Remove unnecessary rows (e.g., repeated headers)
  1. Apply a filter on the column containing sales values
  2. Remove rows with repeated headers caused by merging files
  3. Go to **Transform → Remove Rows → Remove Duplicates**, or filter manually
  • Step 7: Enable automatic refresh
  1. Go to **Data → Query Properties → Refresh data when opening the file**
  2. Optionally set automatic refresh every X minutes
  3. If a new file (e.g., *Sales_Apr.xlsx*) is added to the folder, Power Query will automatically include it upon refresh!

Task

  • ✔ Load and combine data from *Sales_Jan.xlsx*, *Sales_Feb.xlsx*, and *Sales_Mar.xlsx*
  • ✔ Standardize column names and format the data consistently
  • ✔ Remove empty rows, unnecessary columns, and duplicates
  • ✔ Set up auto-refresh so new files are included automatically
  • ✔ Use ChatGPT to optimize the M code in Power Query

Module 7: Optimizing Query Performance in Power Query

Objective

In this module, you will learn:

  • ✔ How to speed up Power Query when working with large datasets
  • ✔ How to avoid inefficient operations that slow down queries
  • ✔ How to use buffering and database-level transformations
  • ✔ How to minimize the amount of data processed for better performance

Introduction

Power Query enables powerful data transformation, but with large datasets, performance can suffer. In this module, you will learn best practices to reduce query execution time.

Instructions

  • Step 1: Avoid unnecessary operations on the entire dataset
  1. Load a large CSV file: 📂Media:PQSales_Large.csv
  2. Check the number of rows and columns – the more data, the more important the optimization
  3. Remove unnecessary columns at the beginning of the query instead of the end
  4. Apply early filtering to reduce the number of rows right after import
  • Step 2: Use buffering (Table.Buffer)
  1. Understand how step-by-step processing works – each operation may cause Power Query to recalculate previous steps
  2. Add `Table.Buffer()` after the filter step to avoid re-processing:
let
Source = Csv.Document(File.Contents("C:\Users\gp\Desktop\PQ\Sales_Large.csv"),[Delimiter=",", Columns=6, Encoding=1252, QuoteStyle=QuoteStyle.None]),
FilteredRows = Table.SelectRows(Source, each [Cost] > 500),
BufferedData = Table.Buffer(FilteredRows)
in
BufferedData


Using `Table.Buffer()` ensures that the results are stored in memory and not recalculated at each step.

  • Step 3: Minimize the number of loaded rows
  1. When working with large databases or CSV files, load only the needed columns and rows
  2. Use **Keep Top Rows** to load e.g., the first 1000 rows for testing
  3. Apply **Remove Duplicates** early to reduce the volume of data being processed
  • Step 4: Optimize database connections
  1. If working with SQL Server, Power BI, or another database, avoid importing full tables into Power Query
  2. Instead, apply filtering and grouping on the database side using native SQL

Example:

let
Source = Sql.Database("ServerName", "DatabaseName", [Query="SELECT OrderID, OrderDate, Customer, Product FROM Sales WHERE Cost > 500"])
in
Source


This ensures Power Query pulls only the filtered data instead of processing the entire table in memory.

  • Step 5: Avoid "drill-down" operations on large datasets
  1. Power Query often suggests drill-downs (e.g., selecting a single value from a table)
  2. When working with large data, operate on whole tables instead of individual records
  • Step 6: Automatically refresh optimized queries
  1. Once optimized, configure the query to refresh regularly
  2. In Excel, go to **Data → Query Properties → Refresh data when opening the file**

Task

  • Load the large CSV file (*PQSales_Large.csv*)
  • Limit the number of loaded rows and columns
  • Apply `Table.Buffer()` and observe performance improvements
  • If using a database, optimize your SQL query
  • Set up auto-refresh for the optimized query
  • Use ChatGPT to analyze performance and further optimize M code

Module 8: Creating Dynamic Reports and Dashboards in Excel with Power Query

Objective

In this module, you will learn:

  • ✔ How to use Power Query to dynamically generate reports
  • ✔ How to combine data from multiple sources into a single report
  • ✔ How to create interactive reports using PivotTables
  • ✔ How to automate report refreshing in Excel

Files Provided

The following files are used for this exercise:

Instructions

Step 1: Import and combine data sources

  1. Open Power Query in Excel
  2. Import the files: *PQ_Sales_Data.xlsx*, *PQ_Regions.xlsx*, and *PQ_Targets.xlsx*
  3. Merge the data using a common key – for example, the `Region` column
  4. Verify that the data is correctly combined and properly formatted

Step 2: Create a dynamic report

  1. Click **Close & Load To...** and select **Pivot Table**
  2. Insert the PivotTable in a new worksheet, using the Power Query output as the source
  3. In the PivotTable Fields pane, set:
 * Rows → `Region`  
 * Columns → `Month`  
 * Values → `Sum of Sales`

Check for accuracy and apply formatting to the table

Step 3: Add conditional formatting

  1. Select the `Sum of Sales` column in the PivotTable
  2. Go to **Conditional Formatting → Color Scales**
  3. Apply gradient colors to highlight low and high sales values
  4. Add a rule: “Greater than” and highlight values above the sales target (from *PQ_Targets.xlsx*) using:
=B5 > VLOOKUP($A5,Targets!$A$2:$B$8,2,0)

Step 4: Automate data refreshing

  1. Go to **Data → Query Properties → Refresh data when opening the file**
  2. Optionally set auto-refresh every X minutes
  3. Test the report by updating the source files and verifying that the report refreshes correctly

Task

  • ✔ Load and combine data from *PQ_Sales_Data.xlsx*, *PQ_Regions.xlsx*, and *PQ_Targets.xlsx*
  • ✔ Create a PivotTable and format it dynamically
  • ✔ Add conditional formatting based on sales targets
  • ✔ Set up automatic data refreshing
  • ✔ Use ChatGPT to analyze and optimize Power Query transformations

Summary: Modules 1–8

Objective

In this exercise, you will summarize all the key concepts learned so far in Power Query by performing a series of transformations on inventory and supplier data. You will apply data import, filtering, merging, column creation, custom functions, and query optimization.

Files Provided

The following files are used for this exercise:

Instructions

🔹 Step 1: Import data

  1. Open Power Query in Excel or Power BI
  2. Import the three CSV files: `PQ_inventory.csv`, `PQ_suppliers.csv`, `PQ_orders.csv`
  3. Make sure all datasets are loaded correctly

🔹 Step 2: Check and convert data types

  1. Ensure all columns in each dataset have correct data types
  2. Issue to solve: the `StockLevel` column was incorrectly imported as text because it contains units like `150 kg`, `200 l`, `75 pcs`
  3. Transform the `StockLevel` column to extract numeric values and store the unit in a new column `Unit`
  4. Verify that the `SupplierID` column is recognized as an integer

🔹 Step 3: Merge data

  1. Merge `PQ_inventory.csv` with `PQ_suppliers.csv` using the `SupplierID` key
  2. Use a **Left Outer Join** to retain all inventory records
  3. Then merge `PQ_orders.csv` with `PQ_inventory.csv` using the `ProductID` key
  4. Verify that supplier and order info have been successfully added to the inventory table

🔹 Step 4: Create custom columns

  1. Add a column `ReorderLevel` that flags products needing restocking when `StockLevel` is less than `MinimumStock`
  2. Add a column `DaysSinceLastOrder` that calculates the number of days since the last order for each product
  3. Create a custom M function that assigns order priority:
if [StockLevel] < [MinimumStock] and [DaysSinceLastOrder] > 30 then "High" 
else if [StockLevel] < [MinimumStock] then "Medium" 
else "Low"
  1. Add a column `OrderPriority` and apply this function

🔹 Step 5: Conditional filtering and transformations

  1. Remove products that have a `Discontinued` status
  2. Add a new column `SupplierRating` that classifies suppliers by reliability:
if [OnTimeDeliveryRate] < 80 then "Excellent" 
else if [OnTimeDeliveryRate] <= 90 then "Good" 
else "Poor"
  1. Verify that the rating logic works correctly

🔹 Step 6: Reshape the data structure

  1. Unpivot columns `Stock_Jan`, `Stock_Feb`, `Stock_Mar` into: `Product`, `Month`, `Stock Level`
  2. Split the `ProductDetails` column into `ProductName` and `Category`
  3. Merge the `SupplierName` and `Country` columns using `" - "` as a separator

🔹 Step 7: Query optimization

  1. Apply `Table.Buffer()` to improve performance
  2. Remove unused columns and duplicates at the beginning of the transformations, not at the end
  3. If working with large data, limit the loaded rows to a test sample of 1000

🔹 Step 8: Export results

  1. Load the final query as a table into Excel
  2. Test data refresh by updating source files
  3. Set up auto-refresh for the query

Task

  • ✔ Complete all the steps listed above
  • ✔ Experiment with both the graphical interface and M code
  • ✔ Apply query optimization to improve performance
  • ✔ Ensure all transformations are correct and results are as expected
  • ✔ Use ChatGPT for troubleshooting or optimizing your M script

Module 9: Importing and Analyzing PDF Files in Power Query

Objective

In this module, you will learn:

  • ✔ How to import data from PDF files into Power Query
  • ✔ How to transform data and perform analysis on business reports
  • ✔ How to visualize results and draw insights from reports

Files Provided

The following PDF reports are used in this exercise:

Instructions

🔹 Task 1: Sales Report Analysis

  1. Calculate total sales for all products
  2. Identify the product with the highest and lowest sales
  3. Compute the average transaction value based on transaction count and total sales
  4. Group data by region and calculate total sales per region
  5. Create a pivot table showing sales by region and product

🔹 Task 2: Employee Attendance Report Analysis

  1. Calculate the average attendance rate across all departments
  2. Identify the department with the highest and lowest attendance
  3. Add a new column classifying attendance into categories:
    1. High: above 95%
    2. Medium: 85%–95%
    3. Low: below 85%
  4. Filter the data to show only employees with low attendance
  5. Create a bar chart showing average attendance by department

🔹 Task 3: Customer Feedback Report Analysis

  1. Calculate the average customer rating on a 1–5 scale
  2. Count how many customers gave a rating of 1 or 5
  3. Generate a summary report of the most frequent positive and negative comments
  4. Sort the data by customer rating from lowest to highest
  5. Create a pie chart showing the distribution of customer ratings

Summary

  • ✔ Complete the analysis tasks for each report separately
  • ✔ Apply filtering, sorting, and grouping operations
  • ✔ Use pivot tables to aggregate data
  • ✔ Visualize results using charts in Excel or Power BI
  • ✔ Use ChatGPT if you encounter difficulties during analysis

Module 10: Importing and Analyzing Web Data in Power Query

Objective

In this module, you will learn:

  • ✔ How to import data from statistical tables available on Wikipedia into Power Query
  • ✔ How to transform and analyze data about countries of the world
  • ✔ How to visualize comparison results in Excel or Power BI

Data Sources

In this exercise, we’ll use real tabular data about countries of the world imported directly from Wikipedia. These include:

  • 📊 **Surface area of countries** 🌍
  • 📊 **Population by country** 👥
  • 📊 **Gross Domestic Product (GDP) by country** 💰

Sources:

Instructions

🔹 Step 1: Import data from Wikipedia

  1. Open Power Query in Excel or Power BI
  2. Choose **Get Data → From Web**
  3. Enter the URL of one of the Wikipedia pages above
  4. Once the available tables are loaded, select the one containing statistical data (e.g., country area, population, or GDP)
  5. Click **Load to Power Query** to begin transforming the data

🔹 Step 2: Transform and clean the data

  1. **Remove unnecessary columns**, keeping only those relevant for analysis
  2. **Change data types** so that numbers are correctly interpreted (e.g., `Area` as number, `GDP` as currency)
  3. **Remove empty values** and correct any errors
  4. **Rename columns** to clearer names, such as `Country`, `Area (km²)`, `Population`, `GDP (billion USD)`

🔹 Step 3: Analyze and compare countries

  1. **Calculate population density** by adding a new column with the formula:
Population Density = Population / Area  
  1. **Sort countries by GDP** to identify the richest and poorest nations
  2. **Compare area and population** to find the largest, smallest, and most populated countries
  3. **Apply filtering** to display only selected continents or world regions

🔹 Step 4: Visualize the results

  1. Create a **pivot table** in Excel to compare area, population, and GDP
  2. Insert a **bar chart** to show the largest economies
  3. Use a **heat map** to illustrate population density by region
  4. Add **conditional formatting** to highlight countries with extreme statistical values

Task

  • ✔ Import world country data from Wikipedia into Power Query
  • ✔ Transform and clean the data to make it analysis-ready
  • ✔ Calculate population density and other statistical indicators
  • ✔ Create charts and comparison tables in Excel or Power BI
  • ✔ Use ChatGPT if you encounter issues during import or analysis