You are not logged in.
Announcement
Unanswered posts
|
- Index
- » Usage, Operation
- » Problem with context variables
Post a reply
Topic review (newest first)
- raghu12345
- 2012-06-05 06:11:35
thank you Djugal
- djugal
- 2012-06-04 17:46:50
you see the user guide mentions that in parallel processing context/global variables does not work.
you have to do a onsubjobok, else you can use a bulk load to load load data faster if needed.
- raghu12345
- 2012-06-04 07:53:32
sorry you are not getting my problem.
I have two jobs named job1 and job2 with the following flow.
job1----------->tmssqlinput---->tjavarow---->(component ok)--->tmssqlinput---->tmap----->tmssqloutput
and i am using
1) In first tmssqlinput i am fetching maximum seq num
2) In tjavarow i am saving the maximum seq value in context variable
3) In tmap i use the same context variable in Numeic.sequnce("S1",context.v1,1)
The job2 is similar to job1 as follows
job2---------->tmssqlinput---->tjavarow---->(component ok)--->tmssqlinput---->tmap----->tmssqloutput
and i am using
1) In first tmssqlinput i am fetching maximum seq num
2) In tjavarow i am saving the maximum seq value in context variable
3) In tmap i use the same context variable in Numeic.sequnce("S1",context.v1,1)
when i run single job at once it will load correctly but I need to run these two jobs at the same time then the problem comes when i run the below job
tparallelize-------(job1)
|
|
------------(job2)
- djugal
- 2012-06-01 14:55:52
it cant work in parallel as the sequence of execution of job 1 has to be completed to fetch the last record
try onsubjobok
- raghu12345
- 2012-06-01 14:18:45
when i run the jobs separately the context variable works fine but when i run the same jobs parallely the sequence number is not coming correctly as required. my job structure looks looks like this
tparallelize-------(job1)
|
|
------------(job2)
and each jobs is like
tmssqlinput---->tjavarow---->(component ok)--->tmssqlinput---->tmap----->tmssqloutput
1) In first tmssqlinput i am fetching maximum seq num
2) In tjavarow i am saving the maximum seq value in context variable
3) In tmap i use the same context variable in Numeic.sequnce("S1",context.v1,1)
I am using same logic in both jobs
- raghu12345
- 2012-05-31 08:52:36
i want sequence number.
after loading first time and when i need to load data again into same tables, it should start the sequence number
from max of sequence number already inserted into table.
for example
for fisrt load
1
2
3
4
and so on upto 1034
for second load it should start from 1035 number
1035
1036
1037
and so on
for saving the previous load max number i am using context variable
It is working fine for one job but i need to run 2 subjobs parallely so i used 2 context variable for each job
then it is mulfunctioning
- janhess
- 2012-05-30 16:35:45
What do you expect? If you want independant sequences use different sequence id.
- raghu12345
- 2012-05-30 14:52:37
i tried to use 2 context variables for fetching max id value of 2 tables,
it is fetching correctly later i am using this max values in 2 sub jobs for sequence number along with data as follows
tmssqlinput---->tjavarow---->(component ok)--->tmssqlinput---->tmap----->tmssqloutput
tmssqlinput---->tjavarow---->(component ok)--->tmssqlinput---->tmap----->tmssqloutput
running these 2 subjobs parallely by using tparallelize component
it is giving sequence no as
1
2
3
4
5
7
9
11
and so on for first table
6
8
10
12
14
16
and so on for second table