Re: Different Number of Synthetic Keys - Qlik Community

igdrazil · ‎2015-11-24

Hi there!

Can someone tell me why i get different number of synthetic keys when i load my script like this:

1. Version:

Here i get 6 syn keys

Inventory_Storage:

LOAD ExpNr,

area,

"buffer",

"TimeStamp" as Zeitpunkt,

"current",

"min",...

JOIN LOAD ExpNr,

area,

"buffer",

"time" as Zeitpunkt,

"min", ...

StatAverage:

LOAD ExpNr,

area,

"buffer",

"time" as Zeitpunkt,

"min",...

JOIN (StatAverage) LOAD ExpNr,

area,

"buffer",

"TimeStamp" as Zeitpunkt,

"current",

"min",...

2. Version:

Here i get 3 Syn Keys

StatAverage:

LOAD ExpNr,

area,

"buffer",

"time" as Zeitpunkt,

"min",...

JOIN (StatAverage) LOAD ExpNr,

area,

"buffer",

"TimeStamp" as Zeitpunkt,

"current",

"min",...

Inventory_Storage:

LOAD ExpNr,

area,

"buffer",

"TimeStamp" as Zeitpunkt,

"current",

"min",...

JOIN LOAD ExpNr,

area,

"buffer",

"time" as Zeitpunkt,

"min", ...

Regards

Chris

EDIT: I removed the SELECT parts and the rest of the Syn-key columns (everything after the ... is identical on all tables)

Anonymous · ‎2015-11-24

Multiple fields are getting repeated in your tables like Area, Buffer, Zeitpunkt etc

You can Rename, Comment, Delete, Composite key of these to avoid...

see these:

What is synthetic keys? How it can be avoided?

Synthetic keys . How to remove them

Synthetic Keys

igdrazil · ‎2015-11-24

Yes i know that, and that wasn't my Question...

My Question was why i get 6 Synthetic Keys when trying the first Version and i get only 3 keys when i load my data with the second Version of the Script.

Regards

Chris

Anonymous · ‎2015-11-24

may be becoz of the sequence of the tables..

Try with same sequence as it is in version1

igdrazil · ‎2015-11-24

Yes but why is the sequence a decisive factor when loading the script?

Regards

Chris

Anonymous · ‎2015-11-24

Not sure, but I would say, mention table name when you are joining..

like

Inventory_Storage:

LOAD ExpNr,

area,

"buffer",

"TimeStamp" as Zeitpunkt,

"current",

"min",...

JOIN LOAD ExpNr, //here and same in version 1 when you join

area,

"buffer",

"time" as Zeitpunkt,

"min", ...

If I am not wrong, you will get same synthetic keys in both the version

igdrazil · ‎2015-11-24

I'll try this for now.

Coming back when finished.

Regards

Chris

raghvendrasingh · ‎2015-11-24

Hi Christoph,

As i think,Order doesn't matter for synthetic keys. order would not effect the model. So could you please share the application with taking no data (take only column as inline). So that we can investigate in better way.

Regards,

Raghvendra

igdrazil · ‎2015-11-24

Just all in all i'm looking for a deeper sense behind this all.

I would like to know why Qlik Sense (and maybe QV) is making different number of synthetic keys while loading the same data with 2 load nearly similar loading scripts (at least the code is the same, just the positioning is different). What is QS trying to do, to think that there should be a differing number of syn keys, so in the long term i may be able to more deeply understand why i should or shouldn't build my script like i do or did.

Thanks for your answers so far. But i'm currently not looking for removing these keys nor making my script better in loading time or something like that . This all was just a curiosity i noticed and hoped i can get some answers to.

Regards

Chris

igdrazil · ‎2015-11-24

I added an example .qvf with marks, which part you need to swap.

In this example qvf you'll get either 1 syn key or 3.

Regards

Chris