Hello, @Agustín Muñoz 

With PQL you can do plenty of operations involving strings. But dealing with free text is always challenging.

Things become easier if you can define in advance a set of keywords that you expect to appear in those comments and which you can use as the basis for analysis.

E.g.:

SUM(CASE WHEN "Your_Table"."Comments_Column" LIKE '%Apple%' THEN "Your_Table"."Revenue" ELSE 0 END)

This statement sums the revenues associated with cases in which the comments include the string "Apple".

I suggest you also learn about the "IN_LIKE" operator and the use of wildcards (%)

Let me know if this was helpful...

See you

Rui Bebiano

Hi @Agustín Muñoz,

Depending on how advanced you want to process free text, you can always try to incorporate the Celonis Machine Learning workbench. It is a Python integration, where tables can be loaded from your data model, processes, and the output can be transferred back to the data model.

Within Python, the options to analyse free text are almost limitless and a lot of free packages exists that can help you with that. You can count the words that are occurring (and filter out stopwords automatically) and do sentiment analyses.

Maybe it is a bridge too far for now, but it's definitely something to take in mind if you want to go deeper in the text analysis field. ;)

Hello @Agustín Muñoz 

For the scenario you mentioned, IN_LIKE operator can be handy and give the desired response.

Remember IN_LIKE can only be applied to STRING values.

Also wildcard usage helps in IN_LIKE operations

The following wildcards can be used in the right pattern expression:

• %: Matches any string with zero or more characters.
• _: Matches exactly one character

Hope it helps.