BWA Attribute Mining

[former_member205400]

I'm trying to understand the mapping of integer ID's in BWA.

My cube is ZPP_C019.

Looking in the TREX attribute mining tab I see as follows (Note that I'm not showing you all the dims here):

PB1_BIC:DZPP_C0191
      Dimid			Value-59	value-117
      sid_0plant		Value-10	Value-
PB1_BIC:FZPP_C019
     $bic$zcount6
     $bic$zcount7
     key_zpp_c0191		Value-59	value-117
     key_zpp_c0192
     key_zpp_c019p
     key_zpp_c019t
     key_zpp_c019u
PB1_BI0:XPLANT
     Sid                                                    
     Plant	
PB1_BI0:SPLANT
     Sid
     Plant	

So I see the mapping correlation between PB1_BIC:DZPP_C0191 Dimid (59, 117) and the fact table PB1_BIC:FZPP_C019 key_zpp_c0191 (59, 117).

What I'm not getting is the following:

1) In the dim PB1_BIC:DZPP_C0191, the dimid (30 rows) and the sid_0plant (24 rows). When I click on each field and then "Distribution" I'm not seeing the correlational mapping. I thought I would see the same number of rows and there would be a clear link between the dimid and the sid_0plant

2) Same kind of question as 1 except that the dim PB1_BIC:DZPP_C0191 sid_0plant should have a direct mapping between the PB1_BI0:SPLANT Sid

3) Also seems to me that the PB1_BI0:SPLANT, the Sid and the Plant should have the same number of rows in it.

Sid has 21 rows and Plant 31 rows.

4) There is a table pb1_zpp_c019 shows all the characteristics and their keys like 0plant_sid and 0plant_key but this seems redundant to the above information

Maybe I'm confused over the purpose of the attribute mining tab and its NOT a window into how BWA has loaded the index (which is really what I wanted to see)

I understand that the dim table is a link between the fact table and the master data I just wanted to go in BWA and verify the mappings.

Any help in my understanding is appreciated.

Mike

[Former Member]

Are you doing this via tcode TREXADMIN or the python tool?

For me the Attribute Mining tab on TREXADMIN is pretty worthless. Try using "Search" or "Mining" in the python tool.

1-4) Distribution - these are distribution of number ranges. The number of rows in the table that it shows when you click "Definition" is not at all the number of SIDS that you have. If your cube has a dimid of 59 for DZPP_C0191 wouldn't that mean you have to have at least 59 rows in that dim table? (unless of course your dim table didnt start at 0)

In the python tool, check out "Search" and then "Search Table". Then you can add stuff like ( dimid=200 ) and you'll see how it matches up with the SID.

[former_member205400]

I'm in the TREXADMIN and don't have access to the stand-alone tool.

I was trying to see the inverted index and how that was all mapped.

Mike