Just an update to keep this going. While I am not yet 100% sure this is the cause I have a suspect. In alot of the cases where Post Updates failed, using the mobilecache2shape tool to extract the modified records has allowed me to notice what seems to be a consistent issue.
I have 8 fields in the feature class to capture numeric test results, the fields are all identical, all are Float with a scale of 3 and a precision of 6 (ex. 123.456). This was based on looking at historic data that was captured previously on paper and then reentered into a non-spatial table. As it turns out, readings can sometimes reach 1000, 1200, etc. What I am seeing is the Posts that fail always have one or two instances of 1000 or 1200, etc in them. Until corrected to say 999.999 I can't even (obviously) load them from shape to the parent SDE FC.
Since noticing this we have educated the users to use 999.999 for readings over that level and the number of failed Posts has dropped to almost nothing. For the two that still occured in the last two weeks, while the user was adament that they were using 999.999 I was able to find one or two higher readings. Lucky for us for any reading that high doesn't really need to be captured as is and can be substituted with 999.999.
If this is the cause I am curious as to how the form in Mobile actually accepts a number with a scale of 4 when the field in the geodatabase is set to 3? Mobile so far has done a great job of honoring what is set up on the desktop, so this anomoly doesn't make sense yet seems to be the cause of our issue.
