Sum based on row value

[Anonymous]

My input is 

Balls,Valid Ball,Runs
1,Y,0
2,Wide Ball,1
3,No Ball,4
4,Y,2
5,Y,6
6,Y,0
7,Wide Ball,1
8,Wide Ball,1
9,Y,4
10,Y,4
11,Y,6
12,Y,0
13,Y,0
14,No Ball,1
15,No Ball,1
16,Wide Ball,4
17,Wide Ball,3
18,Y,6
19,Y,6
20,Y,6

Need Output as

over,runs

1,23

2,33

[TRF]

Do we have to guess the logic for going from the input to the output?

[Anonymous]

oops..forgot to mention that

Over is of 6 Balls, Count 6 Valid Balls for 1 Over.'Y' means its a valid ball

[TRF]

Sorry, I don't play cricket, just play Talend!
Can give more details?

[Anonymous]

need to sum the runs column for 1st 6 'Y's (need to include no ball, wide which comes within 1st 6 valid balls) then for the 2nd 6  'Y's and so on

[TRF]

Here is my proposal:

You need some logic with tMap local variables:

Sum are computed with tAggregateRow component with the following settings:

And the result is:

Starting job test at 11:00 08/10/2019.

[statistics] connecting to socket on port 4014
[statistics] connected
.----+----.
|tLogRow_13|
|=---+---=|
|over|runs|
|=---+---=|
|1   |23  |
|2   |33  |
'----+----'

[statistics] disconnected
Job test ended at 11:00 08/10/2019. [exit code=0]

[Anonymous]

can you please help me with  the logic used in the tmap.

[TRF]

No time for detailed explications.
You have to know local variables are evaluated in the order they are defined.
Start by copying to your job for test then try to decompose manually step by step, it's the better approach to learn (maybe using paper and pen).

Also don't forget to mark your case as solved (Kudos also appreciated).