Understanding data lineage in DAX

[@JohnSullivan1]

Just when I think I am beginning to understand a little DAX I watch one of Alberto's videos and I realize I know nothing. Bravo Alberto! Another great video.

[@albertoferrari6893]

Sorry, but I cannot help replying "You know nothing, John Sullivan!" :) Hope you watched game of thrones, no flames intended, but this sentence circles in my mind since I read your comment :D

[@JohnSullivan1]

@@albertoferrari6893 If only my name was Jon Snow.:) Thanks Alberto. Please keep up the phenomenal work that you, Marco and your staff at SQLBI do.

[@gulhermepereira249]

Same here.. He solves problems that I haven't even came across

[@AdiCristea]

10:51 "and it's actually easier if we try to do that together" - THAT IS what I've been telling my manager all this time, bring Alberto or Marco in, DAX issues will get so much easier when working together; he said no :). Thank you, Alberto, another absolutely great video about DAX!

[@rajunagavelli007]

I really do not find words to say, you are the master of DAX. Kudos

[@cristianosecci7729]

Great video Alberto! Very well explained, and it opens up a lot of possibilities of doing work in a scalable way!

[@mwaltercpa]

Great to know that an expression in a table expression will break lineage, and that TREATAS can restore it. TREATAS is my best friend in a pinch :)

[@mouhcinebouzrar8540]

It's a very deep understanding, thanks for your effort !

[@mwaltercpa]

Thanks for this great video Alberto! I worked through this from a time intelligence perspective wanting to try a list of Calendar month seq numbers shifted back by 1. I called this column “NewSeqNum”. As you point out, this column expression breaks lineage, and can’t be used to *define* filter context… However I am able to call these same new column values within a CALCULATE measure to *modify* filter context. Is this correct? CALCULATE([Sales], FILTER(ALL(Calendar), Calendar[Mo Seq#] = NewSeqNum))

[@DanielADamico]

Every day with new and awesome videos!! Thank you so much for sharing!

[@amarnadhgunakala2901]

That's great vid, now I get what needs to be done for my PBI report.Thanks SQLBI

[@ChetanAdivinnavar]

Great learning.. Thanks much Alberto

[@meloshamefull7492]

Is there a video that will present 10 rows of measures and allow the columns for monthly past totals

[@Elkhamasi]

Now I understand TREATAS after several iterations of reading and watching

[@sqldata6273]

At 11:49 , why the column name in addcolumn() is preceded with "@" e.g. "@FirstSale"? Thanks.

[@SQLBI]

It is explained here: www.sqlbi.com/articles/naming-temporary-columns-in-dax/

[@NyeinChanAung99]

How can we set the data lineage if the table variable has more than one column i.e how to reference that column of table veritable to be treated as a column from the original table.

[@SQLBI]

Use TREATAS with more parameters - see dax.guide/treatas/

[@danc91111]

Does the treatas applied to multiple columns in the model cause in the SE query somehow it to generate a crossjoin to be able to filter the context on the fact table? Given that, if we do the same for two big dimensions (not date as the second dimension), will it cause performance impacts?

[@SQLBI]

TREATAS is managed by the formula engine. You will not see a single SE query with a join, you will see at least two different SE queries, one of them with the list of tuples used in the WHERE condition. Performance might depend on many factors - as usual, measure it in your specific scenario.

[@luisedugarridosoares]

Cio Alberto! Amazing Video! DAX Training courses are online? I mean, not recorded videos, that we can interact with you during the training or we have the both options? I am located in Germany. Thanks!

[@SQLBI]

Our experience is that online training does not work well for interactivity. We tried, but the interaction is minimal and it it way better to use the platform we have, with recorded video training, exercises, and discussion area. The course is cheaper this way and you can invest more in remote consulting if you still need direct assistance.

[@luisedugarridosoares]

@@SQLBI Thanks Alberto!

[@AlbertoGastaldo]

Hello Alberto, one question : when you used TREATAS with the BrandNewNameCorrectLineage, " *NO CONTOSO* " disappeared from the final query results. My question is : the row is not visible because the measure [Amt] for that Brand was BLANK.... or the row is not visible because TREATAS itself removed it from the BrandNewNameCorrectLineage variable as it's not an existing Brand ? My answer would be the first one.... Thanks Alberto

[@AlbertoGastaldo]

Sorry, rows are actually removed by TREATAS itself :-)

[@christianstork1049]

Are there any ways in DAX(Studio) to *explicitly* examine the data lineage of a column? To me, it seems that the didactic problem is that lineage is very much hidden from the user and any way to make it visible would help.

[@SQLBI]

Unfortunately it is not possible without a full DAX Parser, which DAX Studio doesn't have yet.

[@christianstork1049]

@@SQLBI Hmm, do you really mean a DAX *Parser*? I don't see how the lineage is revealed syntactically. Anyway, thanks for the prompt reply! :-)

[@gabrielmorais7312]

NO CONTOSO! Great one, as usual.

[@ClubSoundsForever]

Great, Thank you!

[@peterbogdanovich4043]

Yes, I am enjoying DAX...

[@daniellopez711]

Great video once again! One question, why use VALUES() before certain columns?

[@SQLBI]

It is a way to retrieve data active in filter context.

[@manmeetsinngh1]

Great Video!!! Can you kindly also explain the concept of Expanded Tables with Industrial Standard examples?

[@SQLBI]

See www.sqlbi.com/articles/expanded-tables-in-dax/

[@tangtom2478]

nice video! thank you!

[@felipesignorellireis7839]

Hard for me.

[@SQLBI]

This is a more advanced topic, if you're new to DAX come back to this video in a few months!

[@felipesignorellireis7839]

@@SQLBI Thanks Alberto. There is a lot to learn.

[@cesarsaldana3429]

Mr. Ferrari, he did not know how to explain the causes that break the lineage. It is a pity so many years of experience and he did not know how to transmit it. that was the important thing about this video, restoring it is the least. but what really breaks it?

[@SQLBI]

It's in the video, but if you missed it, you can also find the explanation in the related article: any expression other than a simple column reference breaks the data lineage. www.sqlbi.com/articles/understanding-data-lineage-in-dax/