DSXchange

DS locking is a mostly functional tool to prevent 2 people from editing the same job at the same time. It works for that, more or less -- the biggest issue we have with it is that it can lock you out from the job you were editing (locked by you preventing you from entry) which is mostly caused by ev...

our group managed to connect it into microsoft's TFS but it was too clunky to use. There was some direct connect via a third party library or plug in to tfs. There are all sorts of ways to set something up. you can export jobs/whatever and save those in any normal repository tool. you can also use t...

if you intentionally crash out of, lose the network, or similar while having it open in datastage designer it will lock as you without being actively open (alt f4 will do the trick). However, in either case, be aware that its possible to force them unlocked if someone is annoyed. We wrote a job to d...

it looks to me like you need to add a column to your data for account number, and once you have that, a pivot should work? If you don't want to do a pivot or can't get it to work, you can parse it any number of ways. Datastage can do it with transformers and funnels and such, but it is clunky ... st...

its been a while but if I remember this right, sql server wanted nvarchar not long varchar for long strings. try that?

the first line is inconsistent (it has the column label 3 times, the others only once). Is that correct?

what do you want from it... 3 records? One big record with _1 _2 _3 on the columns?

perhaps less well know, you can call several unix commands back to back with a ; between them in the shell execute stage. here, piped stream is perhaps as good or better, but if you need it... you can invoke the unix commands in a parallel job too, via a routine, if the before/after is insufficient.

it may be easiest to just extract it to a text file, and process the text file in your favorite file fixer -- basic routine? unix shell command? Write your own program? Etc...

if you can't get the transformer, sort by date take top and bottom each down a link and send to a funnel... but the loop is probably the way to go.

1000+ right... personally I would write something that fixed the file outside of datastage for something that big and load the fixed file. But you can probably do it with a transformer loop as you wanted, using the names. If not, a basic routine maybe? that is too many to do it the simple way I said...

you can get the count of records and conditionally bypass the file creation if you are willing to move the parallel job into a sequence job and break it up. Is it worth it?

it looks easy enough. take your input data into a transformer. output 1 of the transformer has this: 11 1 Col1 12 and output #2 has this one: 11 1 Col2 2 put them into a funnel stage. Write to target. I assume there are a fixed number of columns that can be transposed (11,12,13,... N right). each on...

is there a hidden / premium message you need to read in the meantime? Someone here can repost it exposed

"/usr/include/string.h". this concerns me. string.h is part of the C language and should not need a path to include it; has someone tried to override the standard one? try this sample program. Can you compile this? #include <cstring> #include <iostream> using namespace std; int main() { ch...

how complex are these rules, and can you change your extract?
seems like a coalesce or similar statement for each rule would work if the rules are binary (eg column rule 1 = if any of these are null, null else 1) and then trip off the rule columns would be easier if you have this much leeway...