Wrong data type identified in calculations

XaviPolo — Tue, 26 Nov 2024 08:58:33 GMT

When creating new calculations with the FX transformation, in many cases DSP seems to get ‘lost’ in identifying the data type derived from the calculation.
For example the expression

CASE WHEN EMAIL IS NULL THEN true ELSE false END

is correctly identified as a BOOLEAN type.

But with a small modification

CASE WHEN IFNULL(EMAIL, '') = '' THEN true ELSE false END

It gets messed up and says it returns a STRING, which is incorrect.

What do you do in these cases?
Do you force the conversion with something like

TO_BOOLEAN(CASE WHEN IFNULL(EMAIL, '') = '' THEN true ELSE false END)

I'm not sure about the processing impact of having lots of conversion functions (or LEFTs, etc) that are not necessary, but if we don't put them when validating DSP returns the wrong type of data, maybe the HANA optimizer discards them?

Regards,

Re: Wrong data type identified in calculations

Martin_Kuma — Thu, 28 Nov 2024 19:14:41 GMT

DSP messes up the datatypes. I always use CAST("FIELD" AS [datatype]) AS "FIELD", unless you can correct the datatype in the Attributes property - that is easier of course then to CAST

topic Re: Wrong data type identified in calculations in Data and Analytics Forum

Wrong data type identified in calculations

Re: Wrong data type identified in calculations