>>I'm looking for a way to get a relyable checksum/fingerprint of a large character string (2-10MB). The reason why I need such a function is to make sure that we process import files only once. The idea is to add this checksum value to a table. Before processing a new import file, I could then check if the checksum matches a previously processed file.
>>
>>Currently, I'm using a combination of SYS(2007) and LEN(). Anyone know of a more relyable way?
>>
>>TIA
>
>Have you found this method to be unreliable?
>
>Perhaps you should also include the files date and time stamp.
>
>Hilmar.
Depending on the data you import, you might also try a sample import of a few records, and see whether the records already exist. As soon as you detect a duplicate, you can issue TableRevert() or Rollback.
Hilmar.
Difference in opinions hath cost many millions of lives: for instance, whether flesh be bread, or bread be flesh; whether whistling be a vice or a virtue; whether it be better to kiss a post, or throw it into the fire... (from Gulliver's Travels)