EDIReader unable to extract some vendcodes

[Bill Erickson]

Evergreen 2.12

The EDI processing code is able to extract buyer account codes (acq.edi_account.vendcode) from EDI messages when they are appended to the SAN (NAD+BY+<san-value> <vendcode-value>::31B) or vendacct (NAD+BY+...::91) values. This is how B&T and possibly other vendors store vendcode data.

However, buyer account codes may also be stored in a separate field, RFF+API:<vendcode>, following the NAD+BY segment. This variation was seen in the wild w/ Midwest Library Service.

The end result is that inbound edi_messages (ORDRSP, INVOIC) fail to link to the correct edi_account.

EDIReader.pm should be taught to look for codes in the RFF+API and the EDI processing code that attempts to find the correct edi_account for each message should be taught to look at this value first, before attempting to extract the vendcode from the SAN value.

[Bill Erickson (berick)]

http://git.evergreen-ils.org/?p=working/Evergreen.git;a=shortlog;h=refs/heads/user/berick/lp1713138-edireader-vendcodes

I have confirmed the edi_reader.pl extracts the expected results when 0, 1, and 2 RFF+API fields are present. Live environment testing is next.

Also note the only other vendor I've seen that uses RFF+API is Brodart, and in that data they apply a shortened version of the buyer code (and are not also adding the vendcode to the SAN). Since vendcode values never prevent matches, and are only used to improve matches when possible, the code in my branch will be a no-op for Brodart.

[Bill Erickson (berick)]

We've been running this in production for a while. Adding pullrequest.

[Tiffany Little (tslittle)]

We've been running this in production for awhile, so adding my signoff.

I have tested this code and consent to signing off on it with my name, Tiffany Little, and my email address, <email address hidden>.

[Chris Sharp (chrissharp123)]

Pushed to master. Thanks, Bill and Tiffany!