SQL Server Analysis Services Partitions in SSAS Cube

SSAS PARTITION

Real-time Usage:
Multiple Partitions will process the data more parallel.
–> We can run the specified partition to process the required data (so, that limited system resouces will be utilized).
While creating partitions we go for ‘data binding’ between partitions to the table data.
–> There are 2 types of bindings available
a) Table Binding: Here the fact table measure group directly bound to the partition.
b) Query Binding: Here based on the query partition is created.
Note: In real-time we use Query Binding than Table Binding .

Creating Table Binding Partition:

By default a table is bided to the partition in their case the table is FACT TABLE . Along with this FACT TABLE bindings (The cube is having many fact tables). We go for the below process.

Navigation:

Partition –> New partition –> click next.
Measure group: Text Fact
Available Tables: l_l Text _ Fact –>
Click next –> specify the query to restrict rows and remove where condition –>
Next –> processing Location

Current Server Instance:
Storage Location –> click next –> name table-text object partition –> click next finish.
Build –> deploy.

Working With Query Binding:

–> This mechanism we use frequently in real time.
–> Generally we create partitions based on the frequence of data processing and its columns.
–> Assume we are processing the data into fact table based on ‘OK’ and ‘NOK’ flags, then create partiton on table columns.

1) Creating ‘OK’ partition:
(1) Delete the existing table partition.
(2) Partitions –> new partition click –>
Measure group: Text Fact
Available tables l_l Text_Fact
l_l specify a query to restrict rows
SELECT – – – – – – –
– – – – –
– – – –
from [dbo] . [Text_FACT]
Where [dbo] . [Text_FACT] . [IKU] = ‘OK’
Next –>
current server instance –> next
Name: OK_FACt
Aggregation option
Design aggregations later
Finish

2) Like above process create ‘NOK’ partition with below change:

Processing Partitions

There are two ways
(a) Build Menu –> Process
(b) Partitions –> Select partition / partitions –> RC –> Process

Processing FACT TABLE:
If the data (or) structure in fact table changed to effect the same in cube database level we go for fact processing.
Navigation: Partitions tab –> Select Partition –> RC –> Process

Fact Processing Options:
a) Process default
b) Process Full
c) Process Data
d) Process Incremental
e) Process Index
f) Un Process

DIMENSION PROCESSING:

If dimension table structure (or) data changes in data sources to effect the same in cube data base we go for dimension process.
Navigation:
VIEW –> SOLUTION EXPLORER –> DIMENSIONS –> Select Dimension –> RC –> Process
Processing Options:
a) Process default
b) Process Full
c) Process Data
d) Process Index
e) Process Update
f) Un Process

Processing Options For OLAP Objects:
The objects that you can process in SSAS are database, cube, measuregroup, partition, dimension, mining structure and mining model.
Among these objects only dimensions, partitions and mining structures atore data.
–> When you process an object the server creates a processing paln.

Note : ‘Process Add’ is not available in dimension and (fact} partition processing options.
Proactive Cache Practical Implementation :
1. Create a table binding partitions
2. Go To storage settings

Observations:

–> Go To browser –> take some fields and see grand total.
–> add some rows in source “Text_Fact” and after 20 seconds if you see cube automatically proccesses and grand total changes.
Note: No manual intervention.

Understanding Storage Modes:

Enroll for Live Instructor Led MSBI TRAINING

Blog