I'm on 9.4 M4 which seems to be the actual difference. The first time I ran the code I just used the default anydtdtm width just to see which truncates with a default 8 character, and then tried 12 width to account for potential 2-digit hours and that "worked" for me. Stopped too soon. It also works with 11 as a later trial.
I've seen that sometimes that the ANYDT informats need a wider definition than the number of characters/digits actually present. I don't remember the exact case but it involved a datetime value that read the correct value when the Anydtdte used a width of 32 which was several characters wider than the test value. That was arrived at by adding one to the width until finding a width that worked with the particular layout.
It seems this is a bug that has since been fixed then. I can't delete the suggestion.
Interesting that the ANYDT informats exhibit quirky length requirements, I'll remember that. Just like the datetime format. I wonder what's happening behind the scene.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.