Thanks for that. I changed the Number of records on the QSYSPRT file to *nomax and that stopped it halting but it gave the same result.

I will look at the rest.

Rob

As for “creation date/time”, I’m not sure if that’s available anywhere outside of i5/OS-specific APIs, in particular Qp0lGetAttr(). I.e., it’s displayed/printed with DSPLNK, but I suspect that the same or a similar API is used. The Qp0lGetAttr() API isn’t pretty for those who are unfamiliar with that type of API.

In general, if I was doing this, I’d use opendir() to open the directory, I’d read the directory entries with readdir() and use Qp0lGetAttr() to retrieve creation time and then…

Well, actually, I might just use RTVDIRINF for the directory and name some work library to build the information files in. You’ll possibly find the database format of the resulting QAEZD0001O output file easier to work with than either a spooled file or the funky API sequences.

Tom

I totally agree, that is what I am querying.

No, it is a second tier directory that we are reading but all the objects we want to check are in the one level. We get the 200,000 figure from the IFS Catalog created using qsrsrv.

The report I get from DSPLNK is 5838 pages. The first 461 list the short names (giving 461 x 55pp = 25355obj), the next 460 report Long names and the rest are the display attributes. Having looked again I see that each object takes 2 pages to report where my original calculation was made on 2 a page. So It actually only reports full detail on 2500 (approx) objects.

We have tried that method. The problem was that -ctime reports the Last Changed Time and our client wants it based on Creation Time.

Rob

For your example DSPLNK command, I would expect some 9.2 million print lines out of that many streamfiles. Perhaps around half a million pages…

With only 9500 objects, I’d think in terms of 20k pages with maybe 437k print lines. (Give or take some percentage…)

Is this a multi-level directory? If so, do you intend to retain subdirectories?

Tom