Just a quick tip that you might not be aware of in the Power Query Editor.
If you select two columns in the view and on the Add Column tab, select the Date button – you can select to Subtract Days
This will give you the number of days between the dates in the selected columns
Use the formula bar to rename the new column by modifying the step
Hope this can help you too –
Just wanted to share a M function that I just created to calculate the date from a Year and a ISO week number.
The example data is the following
Now the function I created can be called with the Year and Week number as parameters to get the following result
The function has the following code and will return the date of the Monday of the week number.
(TheYear as number, TheWeek as number) as date =>
let
//test
//TheYear = 2018,
//TheWeek = 1,
//
offsetToISO = Date.AddDays(#date(TheYear,1,1),-4),
dayOfWeek = Date.DayOfWeek(offsetToISO),
offset = -dayOfWeek + (TheWeek * 7),
isoWeekDate = Date.AddDays(offsetToISO, offset)
in
isoWeekDate
Hope this can help you too.
Here is a link to an example pbix file – link
Image a scenario where your column in your PowerQuery that contains a table with a set a columns that you know at some point will have more columns added.
In the case above I know that we at some point will add more columns in the merged ProductAttributes table.
When we click the icon for expanding the table, we might just select this and move on
But notice the formula created in
It says
= Table.ExpandTableColumn(#”Merged Queries”, “ProductAttributes”, {“Brand”}, {“Brand”})
This means that even though we might add new columns to the ProductsAttributes table – it will still only be Brand that is expanded and only that column.
The bolded arguments is 2 lists that contains the Column names to expand and the new names of the columns – the last argument is optional so we can actually skip that if we want the original names – https://docs.microsoft.com/en-us/powerquery-m/table-expandtablecolumn
Now by changing the formula to this
= Table.ExpandTableColumn(#”Merged Queries”, “ProductAttributes”,List.RemoveItems(Table.ColumnNames(#”Merged Queries”[ProductAttributes]{0}), {“ProductKey”})
)
We can make the table dynamically expand when adding new columns in the table ProductAttributes
We get the new column included as well
The magic formula does this
Table.ColumnNames(#”Merged Queries”[ProductAttributes]{0})
Will return a list of column names from the step before
expansion (note I use the step name and column name) – and I use the {0} to extract the column names only form the first row – otherwise the formula will fail.
But as we cannot have the same column names twice (i.e. ProductKey needs to go away) so we need to use the List.RemoveItems functions
List.RemoveItems(Table.ColumnNames(#”Merged Queries”[ProductAttributes]{0}), {“ProductKey”})
Thereby removing the ProductKey Item in the list
And this means that when we get more columns in the table “ProductAttributes” table they will automatically be included in the expanded columns
Hope this can help you power queries even more dynamic.
Here is an example file – Link
Power Query On !
Unfortunately I wasn’t able to participate in the Microsoft Business Application Summit this year – but luckily we can watch all the session via https://community.powerbi.com/t5/MBAS-Gallery/bd-p/MBAS_Gallery
But that would mean I had to leave Power BI Desktop in order to search and watch the videos – and the website also made it hard to see the distribution of sessions between categories.
So, I created this –
And to watch the video from within PowerBI I created a drill through page where I used the PowerApp visual to be able to show the
As none of the Microsoft standard visuals can play videos from within a report – I created a power App to show the video based on the selected video link.
If you want to embed this huge resource of learning material in your own tenant you can download the elements from here
The Desktop file – link
The Power App – link
If you are interested in learning how I scraped the web page for all the relevant data – check out these functions to extract data from pages using CSS query capabilities in the power query function Html.Table
Highly inspired by this blog post by Chris Webb – https://blog.crossjoin.co.uk/2018/08/30/power-bi-extract-urls-web-page/
One of the things you should know when working with PowerQuery is that you can get a list of all functions in M by adding a blank query and use the #shared expression to get all the functions.
This can be turned into a table by clicking the “Into Table” button on the ribbon and use the filter to find the function you want to learn about
Now when you click the cell “Value” for a given function the documentation of the function will appear below.
This is almost identical of what you can find online using the https://docs.microsoft.com/en-us/powerquery-m/power-query-m-function-reference – but you won’t have to leave the Query editor
If you use the formula bar and
And complete it without parenthesis
You will get the documentation as well –
This is useful when you write your own Power Query formulas and need more info about the function that the intellisense gives you.
Power ON!
Power Query has a lot of built in functions, but it doesn’t have a function that exactly matches the VLOOKUP in Excel – but by using the M language and other M functions we can certainly replicate some of the VLOOKUP functionality and make an even more flexible lookup function.
Now the example data is like this
In Excel we would lookup the price for at specific productkey by using this formula
– in this case ProductKey 1 with a price of 100.
In order to replicate this in Power Query we can use the function List.PositionOf
So I add a new blank query
And then use the function List.PositionOf – with the following arguments
List – Is the column ProductKey from my lookuptable Products – refer to like this Products[ProductKey]
Value – Is the value to look in this case the value 1
Occurrence – Is set to 0 to only return one value
This will return the position of the value in the list – sort of like using the MATCH function in Excel
Now to return the price – we can use this result to lookup the price like this
= Products[Price]{List.PositionOf(Products[ProductKey], 1, 0)}
And we get 100 returned which is the price of productkey 1.
The formula is structured like this
=NameOfTheTable[NameOfTheColumnToReturnTheValueOf]{PositionReturnedByListPositionOf}
But we why not change it into a function in PowerQuery so we use the function on all rows in a table or on any table.
The function can be created like this
The code
(lookupValue as any, lookupTable as table, lookupColumnName as text, returnColumnValue as text) => let // lookupTable= Products, // lookupColumnName = "ProductKey", // returnColumnValue = "Price", // lookupValue = 1, colLookup = Table.Column(lookupTable, lookupColumnName), colToReturn = Table.Column(lookupTable, returnColumnValue), lookup = List.PositionOf(colLookup, lookupValue, 0), Result = if lookup >=0 then colToReturn{lookup} else "Not found" in Result
The function takes 4 arguments –
lookupValue – The value to find – can be any type
lookupTable – The Table/Query to lookup in
lookupColumnName – The name of the column to lookup the value in
returnColumnValue – The name of the column from the table to return
The colLookup is a variable that uses the function Table.Column to return a list of values in the lookup column.
The colToReturn is a variable that uses the function Table.Column to return a list of values from the values you want to return column.
The lookup variable uses the List.PositionOf to find the position/index of the search value in the lookup column.
Result will use an if statement to test whether a position is found and if so returns the value at the position in the colToReturn list – other wise returns the text “Not Found”.
After this we can use the function in other tables to lookup the for instance the Product price like this – by added a invoke Custom Function Column
OBS – I haven’t tried this on a large table so be aware of any performance issues.
Hope you find this useful – and happy Querying
Here is a link to an example file – Link
I follow EU Eurostat on twitter (link – https://twitter.com/EU_Eurostat ) and often see a lot of interesting facts and infographics like this one.
And I have for a long time wanted to see if I could use the webservices that Eurostat also provides (link – https://ec.europa.eu/eurostat/data/web-services) to import the data directly into Power BI.
So here is a guide on how you can do it – and the example will try to extract the data for the Orange production Infographic.
There is a LOT of different datasets in Eurostat and this guide should work on most of them – you just need to find the dataset (https://ec.europa.eu/eurostat/data/database) in the catalogue.
The REST request we need to construct is defined like this
So, we need to find the datasetCode and specify the filters.
You can find the dataset code by browsing the data catalogue – and the dataset code is stated at the end.
If you need specific items the data explorer you need to specify the code of the items in the request and the Dataexplorer is a good way to find these.
Using the codes we have found we can now use the Query builder to construct our query (click on the picture to try it out)
So after entering the dataset code we get the option to specify the filter and select measures under strucpro
Notice that I have selected Yes to Exclude EU aggregates
The query option will vary from dataset to dataset but the principles are the same.
Clicking the “Generate query filter” will give you
And you can copy the dataset code to clipboard
apro_cpsh1?filterNonGeo=1&precision=1&crops=T1000&strucpro=AR&strucpro=PR_HU_EU&time=2014&time=2015&time=2016&time=2017
Now we have the filter part and this can of course be parametrized in your Power Query.
And we must add
http://ec.europa.eu/eurostat/wdds/rest/data/v2.1/json/en/
before so the full web query is
In order to get the data into Power BI we choose get data and select the Web connector
And press OK
This will return the following to the query editor
We get a JSON response returned from the Web query and this has to be transformed into a table – and in order to get that we need to understand what and how the data is returned.
When we use the query in the browser you can see that its not a normal structured JSON document.
And what we want is a table with the following fields – Country, Year, Area, Production
I start by renaming the query to “QueryResult” and disable the load – and the following query will use this as a reference
So lets create the Geo dimension
By clicking the Record on the dimension field
Drill down to geo
Down to category
And to label
Turn it into a table
And then add an index column starting from 0 and rename the columns
I then do the same for the other dimensions and end up with the following queries
Now in order to get the values we need to do a little more
The web query returns a one long list of values when converted into a table.
Values contains all the “values” in the grid when
Each cell in the table is referred by a running index starting from 0 to Facts multiplied by Geo by Time.
So when we have 2 facts, 38 countries and 4 years will give us 2 * 39 * 4 = 304 – this should be the number of rows in our Values table.
But when we look at the Values before we do anything we have only 282 rows.
The missing Values in because cells with missing values (represented by a : or :z ) is located in the Status field in the QueryResult.
So we have to add the missing rows from this Status – this gives us the missing 22 cells (304-282)
And we then convert these Values to null
In our values Query we want to append these rows to our table – and we can do this by modifying the step – by modifying the expression to
= Record.ToTable(value) & MissingRows
And we rename and change the Name column to Index and change the data type to an integer.
The index column is our number reference to each cell in the table/matrix from EuroStat.
Next step is to calculate the keys for time, geo and Facts.
To calculate the TimeKey we add a column
And divide it by the number to time periods
This gives a step with the following expression
= Table.AddColumn(#”Added Index”, “TimeKey”, each Number.Mod([Index], 4), type number)
And after click OK we can make it more dynamic by modifying the 4 to the number of rows in
= Table.AddColumn(#”Added Index”, “TimeKey”, each Number.Mod([Index], Table.RowCount(Time)), type number)
And now we have the TimeKey for each row.
To add the FactKey we add another calculated column
= Table.AddColumn(#”Inserted Modulo”, “FactKey”, each Number.IntegerDivide([Index], Table.RowCount(Geo)*Table.RowCount(Time)), Int64.Type)
This will give us the factkey and we can see it shifts when we reach row 152 – time count = 4 and geo count = 38 = 152
Now we need the final key column is the geoKey
= Table.AddColumn(#”Inserted Integer-Division”, “GeoKey”, each Number.IntegerDivide([Index], Table.RowCount(Time) ) – [FactKey]*Table.RowCount(Geo), Int64.Type)
And we are now ready to load the data into our data model.
After closing the query editor we get the tables returned to Power BI Desktop and we can finalize the datamodel
And create the relationships between our tables.
And create some measures
And start to visualize the Oranges in Europe
One interesting fact is that the Orange trees are highly productive in Albania.
When 2018 figures is available in EuroStat – we just modify our query to
and refresh our model.
You can download a copy of the pbix file – here.
Did you like it then give the post a like or a comment – let me know what you think – hope you find it useful.
And at least take a spin around Eurostat to what interesting facts that you can find in the many datasets that are provided to you for free.
One of the best functions in Power Query (M) is for sure the function =#shared that inside the Query editor lists all the functions in M and in your current pbix file.
Picture 1 Result of =#shared
But sometimes you do want to check out the documentation online and then you have to leave Power BI desktop – so I thought would it be possible to use Power Query to scrape the documentation (link) – and create a Power BI Report in which I could browse the documentation.
And because the documentation is well structured is turned out to be quite easy
The start page of the Power Query M function reference contains a nice list of all the categories
So begin with we activate a new query with data from the web
And by using the new feature “Add tables using examples” we get a lot of help
Just type in the first 3-4 elements in the “In this section” list and we get a nice result
Click on the edit button and we have a nice list of the categories in our Query editor
Now each of the categories points to a specific web page
And the address for all of them is the name and the hyphens instead of spaces so I do a simple replace values where we add a new column with the relative path
And to get the function category URL I add a parameter called PQ Documentation URL that points to the relative path and that I can use in my other queries as well later.
And then add a column where I concatenate the path and the name of each categories relative path
And now we have a reference to all the pages of the Power Query function categories.
Now to create a table of all the functions I create a new query and reference the first Query I made
And the idea is then to create a custom function that returns a table of all the functions within a category
Each page of contains tables of the functions
And when creating functions I prefer to create a query that works and then turn it into a function so let’s dig in – and the navigator sometimes tries to help us a little bit to much so I rightclick the address and choose edit.
And we get a nice structure of the html page.
We only want the tables and as you can see in the preview each table is actually a list of the functions
Expand the Custom1 column table
And voila – a nice list of all the List functions in this case is returned
The other columns can now be removed, and we are ready to turn this into a function
Open the advanced editor and modify the query from
To
And we can use this function to add as a column in our function query to get all the functions in the different categories
So now we have all the categories and the functions within
Each of the function has a specific page – in this example the List.Accumulate
The address of each of these function pages is the name of the function https://docs.microsoft.com/en-us/powerquery-m/list-accumulate
With a hyphen instead of the . (dot) – so constructing the address is quite easy.
But I only wanted to extract the data within the black square.
By looking at different pages I found out that all of the information I wanted was located in a specific named element <pre>
So on each page I have to locate that tag and then the get all the html for there to a named element called </main> that marks the end of the area.
Now when we use the Web query interface we get a lot of help but in this case we do not want the help
So after clicking Edit in the navigator menu
We get this – and this is because the Web.Contents is wrapped in the Web.Page function
So by modifying it to Text.FromBinary instead we get the html in text returned instead
Now we can click the Fx to create a new step to find the <pre> tag
And another step where we still refer to Source and find the position of </main>
Now as the final step – we extract the html from the source by using the Custom1 and Custom2 value with some offset
And we have the HTML from the page that we wanted.
And we can turn it into a function
Finally I reference the previous query and create the query with all the syntax and example text
Now click close and apply and create a datamodel with the following relationships
I created a measure to the create a default URL to the documentation if a category isn’t selected
And sat the data category to the different URL’s as Web’s url
And the rest is slicers and presentations
To present the HTML extracted from the function page I use the custom visual “HTML Viewer”
And the final report looks like this
And now I can hit refresh and have an updated Power Query Documentation report
I hope you find this useful and if so please like or comment if you have ideas or feedback.
Link to the pbix file is – here
Or if you just want to browse the report – https://app.powerbi.com/view?r=eyJrIjoiMDczMGE5NjMtMjA4Zi00NmE4LTlmODAtZDRhMWQzMTk3ZmUzIiwidCI6ImUwN2Y3YzBiLTk5NWItNDg0NS1iZWJhLTQxNmYwZDYxZTFlMSIsImMiOjh9
A few days ago, I ran into a problem where I had to filter a product table by a dynamically set of columns – so in one solution it could be filtering by field x and y and in other solutions it had to be by field a, b and c.
In this case only rows that where not blank in either of these columns should be included in the result.
Hmmm… how can we do this using Power Query –
Here is the demo data
And if we only want the rows where Field A and Field B has the value “xx” we could add a filter statement (this can only by done if you modify the SelectRows formula manually! )
= Table.SelectRows(#”Changed Type”, each ([Field A] <> “” or [Field B] <> “”))
But what if we wanted to make the fields dynamically
Let’s start by creating a list containing the field names we want to filter
This list could also be retrieved by extracting data from a config file or a database if you would make it even more flexible – just remember it must be at list
Now you should know that any row in a table is a Record
And each record has fields and Power Query has functions related to Records.
https://docs.microsoft.com/en-us/powerquery-m/record-functions
One of this is the Record.SelectFields ( https://docs.microsoft.com/en-us/powerquery-m/record-selectfields ) and this function can be used to select specific fields in a record
So lets add a custom column with this formula
= Table.AddColumn(#”Filtered Rows”, “FilterMultipleColumns”, each Record.SelectFields(_, FieldList))
The first argument uses the _ statement and thereby referring to the current record – and the second is parameter containing the list of fields.
And as you can see when selecting a cell the record in the column now contains only of the fields we have specified in our list.
Now we turn this in to a list by using the function Record.ToList in our AddColumn statement
= Table.AddColumn(#”Filtered Rows”, “FilterMultipleColumns”, each Record.ToList(Record.SelectFields(_, FieldList)))
Notice that we now have a list in the column instead of a record
Next step is to use the function List.MatchesAny to check whether the contains our filter
= Table.AddColumn(#”Changed Type”, “FilterMultipleColumns”, each List.MatchesAny(Record.ToList(Record.SelectFields(_, FieldList)), each _ <> “”))
The second argument is then our filter for all fields and it will return TRUE or FALSE if the criteria is met.
And finally, we can filter the rows using the added Column and only include rows that return true
And if we want to filter by other columns/fields, we change the list
And it works
Hope you find this useful and let me know if you do
Here is a link to a demo pbix .
Query On !
In the Microsoft Flow October 2018 update – link – we got a very interesting feature added to Microsoft Flow – Power Query and as they state –
Also, to better empower our users – we have integrated Power Query directly into the product. Power Query is a data ingestion tool that allows power users to transform, filter, manipulate, and import data into popular data stores like Power BI and CDS. We are adding this same capability to Microsoft Flow that will allow makers to shape data mashups from SQL Server using a Power Query Online experience.
So, I decided to try it out
You can find the Action by searching after “Power Query”
So far, we only have one possible data source– which is the SQL server (but there is more) – so selecting the action gives us this
You can either pick an existing SQL server connection or create a new connection via the dots…
Clicking the Create Query open the Power Query Editor and we can start our transformation/selection of data
And when we click next we get too the Query Editor
Notice that the Server and Database is created as Parameters.
OBS!!!
I haven’t tested it but when you right click the pane you actually get the full “Get data” experience – and when you choose one of these you have to use a Enterprise Gateway to use this.
We can now filter the rows
And even go to the advanced editor when you right click the Query name.
You can only return ONE query to your Flow – so be sure to only enable one of your queries.
When we click the Update button we just get back to our flow
And now we need to add an action that uses the result of the Power Query and we can pick between one of the many possible actions
But just to try let’s try and put the data into Power BI
So, I create a streaming dataset in My Workspace –
And remember to turn the Historical data analysis on
Now as the Query can return more than one row we first Add and “Apply to each” step
And then find the Action “Power BI – Add rows to a dataset”
Now we can pick the workspace – dataset we want the rows add to and bind each column to the result from Power Query
Now we are ready to test our flow
And after the connections is validated we can
Hit continue
And the flow starts
We can follow the progress from the Run history
And entering the step – we can check the output
So, as you can see its JSON document.
And we can start to build a report in Power BI
I think this opens for a lot of different opportunities for utilize the incredible Power Query language to extract and transform data from almost any where and then perform actions – like exporting data, sending notifications via e-mail, create task or alerts based on the data and ….
I am pretty sure that we will have great use of this and can’t wait to see what users will use this for.
Happy Power Querying !!!
Erik Svensen - Blog about Power BI, Power Apps, Power Query 