Synthetics keys

[flygstolen_fred]

Hi,

My app "complain" about use of synthetics keys during the data load process.

Is this possible to solve in a better way and avoid synthetics keys warning? I wanna be able to compare the Reservation data with the Search data and drill down based on geographic's and Supplier. Any help are very appreciated!

Reservation:

LOAD

ReservationId,

Name,

ReservationDate,

TotalPrice,

Supplier,

DestinationCityName;

Search:

LOAD

SearchId,

SearchDate,

Date1,

Date2,

Supplier,

DestinationCityName;

DestinationGEO:

LOAD

    CityName as DestinationCityName,

    CountryCode as DestinationCountryCode,

    CountryName as DestinationCountryName,

    ContinentCode as DestinationContinentCode,

    ContinentName as DestinationContinentName

FROM [lib://Statistics Export Folder/geo.csv]

Message was edited by: fredrik olsson

[Anonymous]

Just noticed you posted a sample, I have updated it to include doing the concatenate, and attached it :

[flygstolen_fred]

Thanks for trying to help Bill. Is concatenation a good way of solving it if the "Reservation table" is connected to approx 20 nested tables but "Search table" is a stand alone table?

[flygstolen_fred]

Or maybe it's better to create some sort of custom key like Vineeth suggested with the values combined.

[Anonymous]

Concatenation is commonly used to create good quality Qlik Data Models.

I would suspect that having approx 20 nested tables is less than ideal.  The most common Qlik Data Model is a Star Schema and it does not sound like you have a Star Schema , but without knowing your complete scenario it is impossible to give a definitive suggestion.

I would generally suggest though that aiming for a Star Schema with a central  Fact Table and satellite Dimension Tables is always a good objective.

Star schema - Wikipedia

[flygstolen_fred]

Hi Bill,

Bad explanation from me, "Reservation" is my main table with 20 tables connected to it in a star schema.

The "Search" table is a more of a stand alone table, but now I wanna compare it.

Reservation table has more than 10 000 entries and Search has more than 1 million. With that amount of entries is it still a good solution to concatenate? And does it matter which table that concatenate the other?

[Anonymous]

Nigh on impossible to guess without knowing the full picture, but a couple of questions :

• What do you mean by The "Search" table is a more of a stand alone table, but now I wanna compare it ?

• What are the 2 fields ReservationId & SearchId used for ?

Maybe helpful if you could share a screen shot of your data model,

[flygstolen_fred]

The Search table only relation to Reservation table is with DestinationCityName & Supplier.

All other tables that have a relation to the Reservation table is with foreign keys.

All Reservation data is actually located in a separate database from the Search data.

The ReservationId and SearchId is only database auto increment numbers.

[Anonymous]

Are the dimensions hanging off the Reservation table relevant to the Search table ?

I'd get rid of the ReservationId and SearchId  if they are not used in the dashboard.

Also I still do not understand what you meant by The "Search" table is a more of a stand alone table, but now I wanna compare it ?

[flygstolen_fred]

Hi Bill,

I think that concatenation is the right way to go in this matter. Thanks for pointing me in the right direction.

Is there any best practice regarding what table that should concatenate the other?