How to tune data loading time in BSO using 14 rules files ?

Question

Hello there,
I'm using Hyperion-Essbase-Admin-Services v11.1.1.2 and the BSO Option.
In a nightly process using MAXL i load new data into one Essbase-cube.
In this nightly update process 14 account-members are updated by running 14 rules files one after another.
These rules files connect 14 times by sql-connection to the same oracle database and the same table.
I use this procedure because i cannot load 2 or more data fields using one rules file.
It takes a long time to load up 14 accounts one after other.
Now my Question: How can I minimise this data loading time ?
This is what I found on Oracle Homepage:
What's New
Oracle Essbase V.11.1.1 Release Highlights
Parallel SQL Data Loads- Supports up to 8 rules files via temporary load buffers.
In an Older Thread John said:
As it is version 11 why not use parallel sql loading, you can specify up to 8 load rules to load data in parallel.
Example:
import database AsoSamp.Sample data
connect as TBC identified by 'password'
using multiple rules_file 'rule1','rule2'
to load_buffer_block starting with buffer_id 100
on error write to "error.txt";
But this is for ASO Option only.
Can I use it in my MAXL also for BSO ?? Is there a sample ?
What else is possible to tune up nightly update time ??
Thanks in advance for every tip,
Zeljko

Answer 1

Thanks a lot for your support. I’m just a little confused.
I will use an example to illustrate my problem a bit more clearly.
This is the basic table, in my case a view, which is queried by all 14 rules files:
column1 --- column2 --- column3 --- column4 --- ... ---column n
dim 1 --- dim 2 --- dim 3 --- data1 --- data2 --- data3 --- ... --- data 14
Region -- ID --- Product --- sales --- cogs ---- discounts --- ... --- amount
West --- D1 --- Coffee --- 11001 --- 1,322 --- 10789 --- ... --- 548
West --- D2 --- Tea10 --- 12011 --- 1,325 --- 10548 --- ... --- 589
West --- S1 --- Tea10 --- 14115 --- 1,699 --- 10145 --- ... --- 852
West --- C3 --- Tea10 --- 21053 --- 1,588 --- 10998 --- ... --- 981
East ---- S2 --- Coffee --- 15563 --- 1,458 --- 10991 --- ... --- 876
East ---- D1 --- Tea10 --- 15894 --- 1,664 --- 11615 --- ... --- 156
East ---- D3 --- Coffee --- 19689 --- 1,989 --- 15615 --- ... --- 986
East ---- C1 --- Coffee --- 18897 --- 1,988 --- 11898 --- ... --- 256
East ---- C3 --- Tea10 --- 11699 --- 1,328 --- 12156 --- ... --- 9896
Following 3 out of 14 (load-) rules files to load the data columns into the cube:
Rules File1:
dim 1 --- dim 2 --- dim 3 --- sales --- ignore --- ignore --- ... --- ignore
Rules File2:
dim 1 --- dim 2 --- dim 3 --- ignore --- cogs --- ignore --- ... --- ignore
Rules File14:
dim 1 --- dim 2 --- dim 3 --- ignore --- ignore --- ignore --- ... --- amount
Is the upper table design what GlennS mentioned as a "Data" column concept which only allows a single numeric data value ?
In this case I cant tag two or more columns as “Data fields”. I just can tag one column as “Data field”. Other data fields I have to tag as “ignore fields during data load”. Otherwise, when I validate the rules file, an Error occurs “only one field can contain the Data Field attribute”.
Or may I skip this error massage and just try to tag all 14 fields as “Data fields” and “load data” ?
Please advise.
Am I right that the other way is to reconstruct the table/view (and the rules files) like follows to load all of the data in one pass:
dim 0 --- dim 1 --- dim 2 --- dim 3 --- data
Account --- Region -- ID --- Product --- data
sales --- West --- D1 --- Coffee --- 11001
sales --- West --- D2 --- Tea10 --- 12011
sales --- West --- S1 --- Tea10 --- 14115
sales --- West --- C3 --- Tea10 --- 21053
sales --- East ---- S2 --- Coffee --- 15563
sales --- East ---- D1 --- Tea10 --- 15894
sales --- East ---- D3 --- Coffee --- 19689
sales --- East ---- C1 --- Coffee --- 18897
sales --- East ---- C3 --- Tea10 --- 11699
cogs --- West --- D1 --- Coffee --- 1,322
cogs --- West --- D2 --- Tea10 --- 1,325
cogs --- West --- S1 --- Tea10 --- 1,699
cogs --- West --- C3 --- Tea10 --- 1,588
cogs --- East ---- S2 --- Coffee --- 1,458
cogs --- East ---- D1 --- Tea10 --- 1,664
cogs --- East ---- D3 --- Coffee --- 1,989
cogs --- East ---- C1 --- Coffee --- 1,988
cogs --- East ---- C3 --- Tea10 --- 1,328
discounts --- West --- D1 --- Coffee --- 10789
discounts --- West --- D2 --- Tea10 --- 10548
discounts --- West --- S1 --- Tea10 --- 10145
discounts --- West --- C3 --- Tea10 --- 10998
discounts --- East ---- S2 --- Coffee --- 10991
discounts --- East ---- D1 --- Tea10 --- 11615
discounts --- East ---- D3 --- Coffee --- 15615
discounts --- East ---- C1 --- Coffee --- 11898
discounts --- East ---- C3 --- Tea10 --- 12156
amount --- West --- D1 --- Coffee --- 548
amount --- West --- D2 --- Tea10 --- 589
amount --- West --- S1 --- Tea10 --- 852
amount --- West --- C3 --- Tea10 --- 981
amount --- East ---- S2 --- Coffee --- 876
amount --- East ---- D1 --- Tea10 --- 156
amount --- East ---- D3 --- Coffee --- 986
amount --- East ---- C1 --- Coffee --- 256
amount --- East ---- C3 --- Tea10 --- 9896
And the third way is to adjust the essbase.cfg parameters DLTHREADSPREPARE and DLTHREADSWRITE (and DLSINGLETHREADPERSTAGE)
I just want to be sure that I understand your suggestions.
Many thanks for awesome help,
Zeljko