Creating a separate column using magic ETL

Stu5677 · 2023-03-28T15:42:32+00:00

There was an error rendering this rich post.

Hi there - I am trying to part of a value that is in a column and change it to be the full name. I keep getting syntax errors whehn going through ETL formula tiles.

My formula is alot larger than this but this is an example - I have tried = instead of Llike, as well as contains, nothing seems to be working (becuase the Campaign column is set up to be WBH - Wabash) so i figured like would be better than =

(Case

when 'Campaign' like 'WBH' then 'Wabash'
Else 'Not Linked'

End)

Find more posts tagged with

Accepted answers

MichelleH

@Stucker A couple notes on your formula:

Campaign is being recognized as a text value instead of a dataset field when using single quotes. Wrapping it in the ` character should resolve this
You will need to include the % wildcard when using the LIKE operator. If WBH is always at the beginning of the string you can use like 'WBH%', or if it can be anywhere in the string you can use like '%WBH%'

All comments

MichelleH

@Stucker A couple notes on your formula:

Campaign is being recognized as a text value instead of a dataset field when using single quotes. Wrapping it in the ` character should resolve this
You will need to include the % wildcard when using the LIKE operator. If WBH is always at the beginning of the string you can use like 'WBH%', or if it can be anywhere in the string you can use like '%WBH%'

JasonAltenburg

I agree with everything @MichelleH said, it's likely the lack of % or difference between ` or ' are causing your syntax errors. I would add to keep in mind that CASE statements are evaluated sequentially when using the LIKE operator.

If your data includes values such as 'WBH - Wisconsin' and 'WBH - Wabash' evaluating like 'WBH%' then 'Wabash' before 'WBH - Wisc%' then 'Wisconsin' will result in both values being evaluated as 'Wabash'

ArborRose

If you are trying to graph or count occurrences in a dynamic way. I would suggest something like this…

Say you create two calculated fields: field_1 and field_2. Field one representing the abbreviation (ie WBH). Field two being the name of the thing (ie. Wabash).

Field_1: SPLIT_PART(Campaign, ' ', 1)
Field_2: SPLIT_PART(Campaign, ' ', 3)

Then just place them in the chart with the aggregate (count in this case).

(apologies…in the actual chart, I used Field3 rather than Field2.)