topic Re: How to put columns and rows properly Tnormalize TUnite in Talend Studio

How to put columns and rows properly Tnormalize TUnite

sakura99 — Wed, 07 Oct 2020 09:29:09 GMT

Halo all, i want put my column sequentially, after using tNormalize and TUnite like this:I already built my component like this:

What went wrong in this case and what should I do?

I really appreciate your time and help. Thank you.

Re: How to put columns and rows properly Tnormalize TUnite

manodwhb — Wed, 07 Oct 2020 09:51:09 GMT

@hallo sakura , you can tAggregarerow and you can use max of value for each coulumn and based on Id and you can do that way.

Thanks,

Manohar

Re: How to put columns and rows properly Tnormalize TUnite

Anonymous — Wed, 07 Oct 2020 10:12:36 GMT

Hey,

I would extract the columns in three different stages, then extract the values (based on the "|" delimiter (discard empty rows) and finally perform a join between your different outcomes (extracts). ==> This solution does not meet your requirement, see below question:

What is important (and that is something that I don't understand right now): how do you determine wheter the value "AA" is assigned to "50000" and not "-90"?

The amount of pipes between the columns does not match, so is there any logic used? (it seems to be first value of every column is one match, and so on...)

I hope my question is clear, if not, don't hesitate to answer on this topic.

Kind regards,

JVL

Re: How to put columns and rows properly Tnormalize TUnite

sakura99 — Wed, 07 Oct 2020 14:01:58 GMT

ah i forgot.

I can write like this:

Extract if any values between the pipelines, (for each column)

Transpose it (for each column)

Unite it

Filter by EHKOTAK columns. EHKOTAK columns must be not null. The others can be null.

Re: How to put columns and rows properly Tnormalize TUnite

DataTeam1 — Wed, 07 Oct 2020 18:19:42 GMT

@hallo sakura after using tNormalize on 3 columns you should get 3 tables. Every of it should have 3 columns: ID, row_number, value.

To create row_number you should use Numeric.sequence("key",0,1) where key is different for every column. You can use i.e. EHKOTAK, EHPERSEGI, EHBULAT as keys names.

Then you should use tMap (instead of tUnite) to join (lookup) 3 tables on ID and row_number columns