named pipes

[kollurianu]

i would like know the purpose of name pipes sequential stage , i read the
help , did not get and an how to use it or where it exactly used.

any examples appreciated.

Thank you all.

[ArndW]

Kollu,

a named pipe is rather like a sequential file, with the exception that the data is never actually written to disk. You have one writer and one reader (which are usually different processes) and a buffer which is in actuality the "pipe". Since the data is only found in memory it is not persistent and both processes get read or write errors when the other is no longer attached to the pipe.

There are a lot of places where pipes are used to great enhancements in speed. Just think of two separate DS jobs, where the output of one is used as the input of another; and the data is never needed again. Assuming 1 million rows using a table or sequential file would incur 2 million + Disk I/O operations whilst changing that to a pipe would just mean that the CPU has to process the actual data amount plus a bit of overhead for buffer management. The math is easy to do in this case.

[Sainath.Srinivasan]

You need to create the pipe before using it.

[kollurianu]

how to create a pipe? any example jobs of named pipes ., iam not getting
clear picture of how this going to work.

Thank you.

[ray.wurlod]

Depending on your UNIX, a named pipe is created with a command such as mknod or mkfifo - check with your UNIX man pages.

(There's a lot of peripheral stuff you need to know to be a DataStage consultant, isn't there?!!)

[kollurianu]

Yes there is alot of peripheral stuff to be know Ray , thanks a lot for ur reply.

Thank you all,