count of the column by using loop variables

[iskapalli]

Hi,

I want to caclulate count of the column by useing loop variables in transformer.

select count(col1) from the table.

But i am getting 5000 rows out of 10000.


Defined the following stage variables:

NumSavedRows
SaveInputRecord()

IsBreak
LastRow()

NumRows
IF IsBreak THEN NumSavedRows ELSE 0


Defined the following loop variable:

SavedRowIndex
GetSavedInputRecord()

Loop condition:
@ITERATION <= NumRows

Define the output link columns and their derivations:
Col1 - NumRows

[ray.wurlod]

Anything in the parallel Transformer stage is per-node.

[Kryt0n]

I want to caclulate count of the column by useing loop variables in transformer.

select count(col1) from the table.

Why would you need to advance passed this if all you want is the count?

Defined the following stage variables:

NumSavedRows
SaveInputRecord()

Define the output link columns and their derivations:
Col1 - NumRows

If you aren't using the input record, why save it?
Is there a requirement behind what you are doing or practising the use of looping?

[iskapalli]

this is the requirement. This is the first time iam implementing by useing loop variable.I am trying to give lastrow(col1).but it is not allowing. Please let me know where I am doing wrong

[Kryt0n]

So you only want the last row of data? Just have a LastRow() condition on the output stream, run in sequential mode and sort your input... or better than that, make sure you only retrieve the row you want.

Looping is only required when you want one row to produce multiple rows so unless there is something missing in your requirement, I don't see what it is.

[iskapalli]

Sorry Kryt0n I will explain my requirement clearly.

job design-> td stage -> transformer-> td stage

in source I have written custom query. From this query I am extracting one
column(period). I want count of this column. For this I want to use loop variable. I hope you will get my requirement.

[chulett]

Not really clear yet, I'm afraid. Can you post an example or twelve? What the selected data looks like and what your counted output from that input would be? That would be a good and much less ambiguous start.

[PhilHibbs]

Loop variables are not for counting the number of rows - Loops replicate the input row a number of times on the output link, and Loop Variables are evaluated for each copy of the input row that is generated.

To count the number of rows, feed your stream through an Aggregator.

And if you really want ot do it with a Loop Variable... well, don't. That's not a sensible thing to want to do. It's like wanting to drive across a river. Sure, if you go fast enough you might make it, but the only sane solution is to use a bridge or a ferry.

[ray.wurlod]

You could also wait for a drought.

Meantime you can count in a Transformer stage using stage variables.
As noted, Loop variables are not the correct mechanism. Use a Loop variable for processing things within a field (though counts of those can probably more easily be obtained with functions like Count() or DCount()).

[chulett]

Or we could wait for an example of the input and desired output so we have a better idea what "count" means in this context...

[iskapalli]

example

input:
col1
a
b
c
d
e
f

output:
col1 col2
a 6
b 6
c 6
d 6
e 6
f 6

please help me

[ray.wurlod]

Search DSXchange for "fork join". This is not a task for the Transformer stage.