Sort option

[dslearner07]

Hello All

I am new to datastage though following this forum for some time now. I have a simple question on Sort stage. Appreciate if someone answer with a simple example. Only example will be enough to show the difference. I have gone through the forum history but bit confused with all answer.

"What is the differerence between the Cluster key change and key change option?"

Thanks

[ray.wurlod]

Cluster is where the rows are clustered (contiguous) but not necessarily sorted.

[dslearner07]

If i have below input data

Col1 Col2
10 Delhi
10 Mumbai
30 Adelade
40 NewYork
50 Sydney
50 LA
50 HK

What would be output of above for both clustered and key change? Assume both cluster and sort is on Col1. Appriciate your resposnse.

[Satwika]

Hi ,

According to my understanding :

For Cluster key column-- sort stage itself creats the clusterkeychange in output column , and takes the default 'tinyint ' datatype. The 'sort key mode' should be in either 'Don't Sort(previously sorted/Previously grouped mode'. Cluster key mode is not ready to take the sort option.

Both the Cluster key , Key column change gives the same output as below.

In Key column change it will sort the data in specified mode and it will create the keycolumn change column in output.

For your input ,

Col1 Col2
10 Delhi
10 Mumbai
30 Adelade
40 NewYork
50 Sydney
50 LA
50 HK


Expected output :

Col1 Col2 Clusterkey
10 Delhi 1
10 Mumbai 0
30 Adelade 1
40 NewYork 1
50 Sydney 1
50 LA 0
50 HK 0

Regards
Satwika

[ray.wurlod]

If there's only one key column, then cluster key and sort key tend to be the same. In your example, Col1=50 forms a cluster even if Col2 is not a sort key.

[dslearner07]

Thanks Satwika.. Can you just put an example where both Cluster and key change column output would be different.

[Satwika]

Hi,

Try with the 2 key columns define in your input and make a sort on one key column and perform cluster on another key column it will give the different output.