help in logic

[pandeesh]

Hi,
My source sequential file contains 4 fields:

Code: Select all

id,division,date,amount
--------------------------
1,0,10-10-2011,1
1,0,11-10-2011,12
1,0,12-10-2011,23
1,1,1-1-2011,12
1,1,2-1-2011,13

My expected output should be like below:

Code: Select all

id,division,date,amount,amount1
-------------------------------------
1,0,10-10-2011,1,1
1,0,11-10-2011,12,13
1,0,12-10-2011,23,36
1,1,1-1-2011,12,12
1,1,2-1-2011,13,25

I have achieved this using the below design:

Code: Select all

seqfile------linksort on id,division and date--->Sortstage(id,division)-->xfm-->xfm1--->tgt_dataset

i am using two transformer stages. In the first transformer, i am introducing the column amount1 with amount values for the first record in each group.
For the first record in each group the amount1 will be same as amount field.
The remaining records will have 0 in amount1.

in the second transformer, i am calculating the amount1 using stage variables.

Is it possible to achieve the same using a single transformer stage

Thanks

[ray.wurlod]

Use stage variables (and maybe a Sort stage generating key change column sorting on id and division). Increment counter if no change, reset counter if there is a change.

[pandeesh]

in my design, i am using 2 transformers xfm and xfm1.In XFM, i am adding a new column Amount1 with below values:

1)In each group ,for the first record (whose Keychangecolumn is 1), i am populating amount value as it is to amount1.

2)For the remaining records in the group i am setting 0.

In xfm1(2nd transformer stage),i am summing up amount1 with amount as per the requirement,
The stage variables i am using are:

Code: Select all

StageVar=Amount1
StageVar1=if keychangecolumn=1 then StageVar else StageVar2+Amount
StageVar2=StageVar1

And passing StageVar1 to Amount1 field.
i am able to get the desired output.But i am using 2 transformer stages(xfm,xfm1).

Is there any other way to achieve this using a single transformer?

Thanks

[chulett]

I don't see anything about this that requires two transformers. It's just a running subtotal and you can do all that in one transformer along with the 'group change detection' you are doing.

[pandeesh]

My understanding is:

1)The input to the first transformer doesn't have amount1 fields .it has only amount field only.

2)In the transformer i can hold the value of the previous input record. i can hold the previous value of amount but not amount1,

As per my requirement, i need to add the Previous value of amount1 with current value of amount.

i dont know how to maintain the summed up values of amount(in stead of considering amount1) in stage variable.

Please help me.

Thanks

[kamtammystuff]

I see your problem here, U can use the same logic you mentiuoned in a single transformer and before that you need to initialze the StageVar variable to 0. So that for the very first record you will have 0 + Amount and then subsequently you'll be able to add Amount to StageVar1.

[pandeesh]

I will try that:

Code: Select all

StageVar1=if keychange=1 then amount else amount+StageVar2
StageVar2=StageVar1

Initial Value:
StageVar1=0

[ray.wurlod]

U can use the same logic ...

... but do you think U (one of our posters) would actually need to?

The second person personal pronoun in English is spelled "you", not "u".

[ray.wurlod]

Every If needs a Then (as well as an Else) in DataStage expressions.

[pandeesh]

Edited the expression with then clause

[chulett]

And?

[pandeesh]

It's working fine