Renaming All Tables’ Columns in One Go in Energy Question

I beforehand wrote a weblog submit explaining easy methods to rename all columns in a desk in a single go together with Energy Question. Considered one of my guests raised a query within the feedback concerning the chance to rename all columns from all tables in a single go. Curiously sufficient, considered one of my prospects had the same requirement. So I believed it’s good to put in writing a Fast Tip explaining easy methods to meet the requirement.

The Downside

You’re connecting to the information sources from Energy BI Desktop (or Excel or Information Flows). The columns of the supply tables are usually not consumer pleasant, so that you require to rename all columns. You already know easy methods to rename all columns of a desk in a single go however you’d like to use the renaming columns patterns to all tables.

The Resolution

The answer is sort of easy. We require to hook up with the supply, however we don’t navigate to any tables immediately. In my case, my supply desk is an on-premises SQL Server. So I hook up with the SQL Server occasion utilizing the Sql.Database(Server, DB) perform in Energy Question the place the Server and the DB are question parameters. Learn extra about question parameters right here. The outcomes would really like the next picture:

As you see within the above picture, the outcomes embody Tables, Views and Features. We’re not fascinated with Features due to this fact we simply filter them out. The next picture exhibits the outcomes after making use of the filter:

If we glance nearer to the Information column, we see that the column is certainly a Structured Column. The structured values of the Information column are Desk values. If we click on on a cell (not on the Desk worth of the cell), we will see the precise underlying knowledge, as proven within the following picture:

Because the above picture illustrates, the chosen cell incorporates the precise knowledge of the DimProduct desk from the supply. What we’re after is to rename all columns from all tables. So we will use the Desk.TransformColumnNames(desk as desk, NameGenerator as perform) perform to rename all tables’ columns. We have to move the values of the Information column to the desk operand of the Desk.TransformColumnNames() perform. The second operand of the Desk.TransformColumnNames() perform requires a perform to generate the names. In my instance, the column names are CamelCased. So the NameGenerator perform should rework a column title like EnglishProductName to English Product Identify. As you see, I want to separate the column title when the characters transit from decrease case to higher case. I can obtain this through the use of the Splitter.SplitTextByCharacterTransition(earlier than as anynonnull, after as anynonnull) perform. So the expression to separate the column names based mostly on their character transition from decrease case to higher case appears to be like like under:

Splitter.SplitTextByCharacterTransition({"a".."z"}, {"A".."Z"})

As per the documentation , the Splitter.SplitTextByCharacterTransition() perform returns a perform that splits a textual content into an inventory of textual content. So the next expression is reliable:

Splitter.SplitTextByCharacterTransition({"a".."z"}, {"A".."Z"})("EnglishProductName")

The next picture exhibits the outcomes of the above expression:

However what I want just isn’t an inventory, I want a textual content that mixes the values of the listing separated by an area character. Such a textual content can be utilized for the column names. So I exploit the Textual content.Mix(texts as listing, non-obligatory separator as nullable textual content) perform to get the specified outcome. So my expression appears to be like like under:

Textual content.Mix(
            Splitter.SplitTextByCharacterTransition({"a".."z"}, {"A".."Z"})("EnglishProductName")
            , " "
            )

Right here is the results of the above expression:

So, we will now use the latter expression because the NameGenerator operand of the Desk.TransformColumnNames() perform with a minor modification; reasonably than a relentless textual content we have to move the column names to the Desk.TransformColumnNames() perform. The ultimate expression appears to be like like this:

Desk.TransformColumnNames(
                    [Data]
                    , (OldColumnNames) => 
                    Textual content.Mix(Splitter.SplitTextByCharacterTransition({"a".."z"}, {"A".."Z"})(OldColumnNames)
                    , " ")
                    )

Now we will add a Customized Column with the previous expression as proven within the picture under:

The next picture exhibits the contents of the DimProduct desk with renamed columns:

The final piece of the puzzle is to navigate by the tables. It is vitally easy, good click on on a cell from the Columns Renamed column and click on Add as a New Question from the context menu as proven within the following picture:

And… right here is the outcome:

Does it Fold?

That is certainly a basic query that it’s essential to at all times ask when coping with the information sources that assist Question Folding. And… the fast reply to that query is, sure it does. The next picture exhibits the native question handed to the back-end knowledge supply by right-clicking the final step and clicking View Native Question:

If you’re not aware of the time period “Question Folding”, I encourage you to be taught extra about it. Listed below are some good sources:

Conclusion

As you see, we will use this method to rename all tables’ columns in a single base question. We should always disable the question’s knowledge load as we don’t have to load it into the information mannequin. However remember, we nonetheless have to increase each single desk as a brand new question by right-clicking on every cell of the Columns Renamed column and choosing Add as a New Question from the context menu. The opposite level to notice is that everybody’s circumstances might be completely different. In my case the column names are in CamelCase, this may be very completely different in your case. So I don’t declare that we totally automated the entire strategy of renaming tables’ columns and navigating the tables. The desk navigation half remains to be a bit laborious, however this method can save a whole lot of growth time.

As at all times, when you’ve got a greater concept I admire should you can share it with us within the feedback part under.

You’ll be able to obtain the pattern PBIX file from right here.