SID (Master Data)

[Aummad]

Hi,

Please some one explaine me how the SID working with example,

specially if the master data not available,

suppose we load the TD into the ODS and cube without master data,

MS will load through the SID.

Please help me to understand this concept.

Thanks,

[Former Member]

Hi Rob,

(surrogate ID: an external ID for master data you load. For example if you have 1000 customers in your system, but your transaction data has 100 different customers, so the system will give SID's to only those 100 customers if you do not load master data but load TD)

Assume that you are loading sales data which includes customer, material, date and quantity and unit.

Here is a data:

C100 , M100, 01012008, 100, KG

C500 , M100, 01012008, 30, KG

If you have no data in your cube and master data for customer and material,

System checks if C100 is previously loaded for customer master data. Answer is NO for our case.

So creates and SID lets say '1' for C100.

The SID table will have a data like follows.

SIDforCustomer :1 Customer : C100

For C200, check again the master data if it exist in SID, Answer is NO for our case. Create another SID for C500.

SIDforCustomer :1, Customer : C100

SIDforCustomer : 2, Customer: C500

Same rule applies for material and all master data carrying objects.

If you do not load master data before transactional data, you will have only SID's and Object Key values in your master data tables ( like P, Q ...) But the attribute values will not be loaded as you can expect.

In most cases, it is a must to load master data before TD.

Like, if you read some attributes from some master data.

Hope this helps.

Derya

[Aummad]

Hi,

thanks for reply,

so it's mean key values creating through SID are valid values.

suppose material, sold_to, ship_to are key fields in the characteristic, when the TD come, it will creat the valid materil, sold_to, ship_to values in the coresponding tables,

is this right.

Thanks

[Former Member]

Hi Rob,

If you have a characteristic with key fields material, sold_to, ship_to , and you put this characteristic in a cube or ODS, the compounding attibutes like sold_to, ship_to will be automatically added to the cube or ODS. ( you can not remove them )

Meaning that, your TD must also have those information in it. Again meaning that, your SID generations will be fine. Nothing to worry, you can test this case easily with a simple cube and character you mention .

But my little advice would be, load master data before transactional...

Hope this helps...

[Former Member]

Dear,

1. SID links the Master Data to the DIM tables.

2. Imagine master data is already loaded. Now, when you load the transaction data, it matches with the corresponding key in the master data and creates a SID.

3. Imagine if the master data is loaded afetr the transaction data. In this case, the look up becomes reverse, which means, for every master data that is getting loaded, it needs to scan through the entire transaction data to create SIDs. (This is the reason to load master data first and then load transaction data after that).

4. If no master data is present, then, when the transaction data is loaded, it just creates an entry for the master data in the master data table without any text/attributes and the SIDs are also created(since, master data is not loaded). In this case, if you check the contents of the master data table, you can find that the primary keys are loaded, eventhough the master data is not loaded, but the attributes and texts are not loaded in the master data table.)

Thanks

[Former Member]

Hi,

SID stands for "Surrogate ID"

For every Master/Transaction data you load to the cubes, the system automatically generates SID to represent data, because it is easier & faster to process Numeric characters rather than Alpha-numeric.

Even though you may get data for an infoobject by this but you wouldn't get the data for the attributes...

Hope it helps,

Ajay