DSXchange

Sorry, but that's not what I said. Your timestamp has has a precision of three so the mask needs to be %ss.3 rather than %ss.6.

It's %mmm rather than %mon from what I recall and you'll need six zeroes for %ss.6 and you've only got three.

One should be able to make that work as well, even though it violates the "without any command in unix" rule. Technically.

Just svreject could work. But for this to make sense, svreject as "false" [ aka Not(svreject) ] would be used in your target link, while svreject as "true" would be for a reject link.

No clue.

What 8.x version? Are you current on patches / fixpacks? Any chance that Master Sequence's log is quite... large?

Has this run before? If not, I would check your path for permission issues, ensure you have the rights you need to create a file there.

What stage and 'action' are you using?

ps. That array size will kill your performance, it should be something larger - for lack of knowledge and thus any better suggestion try 1000 and see if that helps.

Yes, but that would be 'too late' in this case. Well... unless you wrote everything to the file, got the count after job and then overwrote the file with the total. I suppose.

Or as Ray noted, read it as a varchar and use the appropriate mask with the TO_DATE() function in your target sql.

Of course, another option is to get whomever or whatever is producing that file to build a valid csv file rather than junk like that.

Use a lookup so you know both if it is "new" or not and what the old values are. Then you can either insert or update and increment the old values appropriately.

No. Perhaps in 8.5 when it can recognize EOD or perhaps with looping but older versions will require aggregation. Or just use a script to get the line count and echo it to the output file.

This is all 'Command Line 101' stuff. :? You can run any command from anywhere as long as the directory where it lives is in your PATH environment variable. When it isn't you have to tell the O/S where the command can be found and you can do that in two ways: 1. Use the full path to the command as T...

'Sweet spot' means you have to try various values, dialing it up and back down again until you find the appropriate value for this combination. Many things can affect this with the setting you've mentioned only part of it, for example your average record length, indexes on the target, disk subsystem...

Or a comma space, or a semi-colon space - some combination should work. And there's no way to "cc" anyone out of the box, you'd have to write your own mechanism to do that.