How to Update a Data Source in Power BI (without losing all of your work)

[@bigmo931]

This is gold!!! This was extremely helpful. I don't know how many times I had to change data sources for one reason or another and I had to basically reinvent the wheel with the change steps and relationships.

[@davidterry6312]

This was top notch video - thank you!

[@txreal2]

I learned the hard way to put each data source in its separate query (and duplicate or reference before modifying Applied Steps) then use them as "Source" for other queries. Modulize your queries and re-use them or save as templates. Also make data file path dynamically by creating in the current workbook: folder path, file name, and CONCAT then to make file path name ranges; then import file path in a separate query step.

[@rowanschoultz1022]

Awesome tip, thank you!

[@elrevesyelderecho]

Thanks for this video Question. This work just for 1 PBI report. What can I do having several PBI reports? All PBI reports are connected to the same data source/Server location and I need to change it to a new Server location. Is there a way to change all data source across all PBI reports rather than open and do it one by one? Thanks

[@mrbartuss1]

You can even set parameters to switch between the sources

[@abdelazizsaafane190]

Thank you !

[@hemanthkumar2704]

Great video. I tried this and encountered an error with changed type step in power query editor. Everything was fine until this error. This is the most common error it seems. Any comments?

[@dotnetdevni]

For some reason when I added a new table it seems to do a refresh and cause the table to not be in the model anymore no idea why

[@jennyzhang8266]

This is awesome!!!

[@seaprincesshnb]

Great video. Please stop hitting your desk - it causes a low boom in my headphones that sounds like an earthquake.

[@TSmitty360]

I have new dataset, same fields, lowercase headers now instead of upper case, does that matter?

[@Nick_Lee_PW]

Yes, unfortunately it does matter. The M code in Power Query that you're going to be modifying is case sensitive, in contrast to the DAX language in PBI Desktop is not case sensitive. I would recommend updating your original sources to match before performing this method.

[@sonjoysengupto]

Will this work if I migrate from a Cloudera source to a Snowflake source?

[@joachimkober9941]

This is great. BUT what if the data source is changing all headers to small letters? And before the headers were written in camel (capital and small)? This breaks my relations and all power query code. Do you have a solution for this?

[@leonid123ful]

No solution to this. You must make sure all columns are exactly the same.

[@Nick_Lee_PW]

You would have to modify the M code a little bit. If it's just a column/header name difference, you can just change any reference to those column headers in the M code directly.

[@txreal2]

There's more than 1 way to do this. Since csv works faster in Excel Power Query, I do this all the time for same usually large data table. I would delete the 1st row of the csv table as a data source. Open the Excel header table (import it or put in current workbook) then merge cvs table into it. That's it. There are other ways if you KZbin search.

[@wozturner6999]

Unfortunately it only works if the fields and table structure are like for like. I have scenarios where users create an excel flatfile without any optimisation ...builds a BI report and says can you put that in the service app. NO!!! I then have to create a star schema (where applicable), remove cardinality, create explicit measures rather than implicit, optimise the model and all of the visuals are effectively broken as a result, so I then have to rebuid everything more or less from scratch before it gets approved and promoted to an app. If anyone has a non manual solution to those scenarios please let me know and yes training our users is what Im already trying to do...but explaining do it this way, when they can do it that way is the excuse and is my issue not theirs in their eyes ..GRRR.

[@fdtank81]

You gotta be kidding me, this makes no sense

[@siddiquisherry7483]

So helpful, thank you ☺️