Synthetic Keys

[rittermd]

Let's say that you write a script with multiple sections.  You load 1 table and then a 2nd.  If you only run these 2 sections you have generated a synthetic key with 4 fields. 

If you let the rest of the script run by the time you get to the end you have no synthetic keys because tables have been dropped along the way.

My question is, does this matter?  Are synthetic keys only created at the end of the script?  Or do you want to make sure that you don't have any from start to finish?

I hope that made sense.

[Vegar]

You do not need to worry about synthetic keys during script execution.

[#3 · Accepted solution]

For good practices you should not have any synthetic keys

[OmarBenSalem]

That's a really good question !
stalwar1‌, your response on this matter would be much appreciated. thanks

[Vegar]

All synthetic keys are generated straight after the the end of your script execution. During the script execution there are no synthetic link tables in your data.

[rittermd]

Just to confirm then.  It does not matter that tables in step 1 could associate with 4 fields if by the time the script ends that these associations are removed?  This does not affect the processing of the data?

[klausskalts]

That is corect - it's only at the end of the script-execution the datamodel is finally build and the connections (and syjthetic keys are created.).