wierdness when importing contacts with the REST API

[ljm]

Hi, I'm seeing an odd thing when using the REST API to import contacts.

The attached contacts.csv.txt (.txt suffix so that it be attached to this forum thread) contains 20 rows plus a header row.

When I import it via:
curl -u userassword --data-binary @/tmp/contacts.csv.txt \
http://zimbra.example.com/service/ho...ntacts?fmt=csv

I only see the first 16 contacts inside zimbra. That's the case whether I log i via the web-based UI, or suck them out again using http GET. I can only see 16 "blah ... mailbox - UserServlet (POST) ... blah" lines in /opt/zimbra/log/zimbra.log Stopping and starting zimbra doesn't help.

But, if I split this one file into 20 files, each with one line of header and one line of contact, and import each file separately:
for i in contact-part-*; do
curl -u userassword --data-binary @/tmp/$i \
http://zimbra.example.com/service/ho...ntacts?fmt=csv
done
it works as expected. 20 contacts in zimbra.

I originally began working with 500 contact lines in a .csv, where zimbra took the first 16, skipped some, took a few more and so on. Out of 500-odd contacts it imported maybe 30-odd.

I'm happy with the workaround, just thought I'd post for interest.

[schemers]

I noticed that some lines ended with:

...-home-web-page,

and (5 to be exact) did not:

...-home-web-page

If I added a ',' to the end of all them then all 20 import in one shot.

I need to do a little more research to figure out if it is a bug or not, but it sounds like it might be.

Can you please a bug report?

thanks.

[ljm]

Looks like when importing a .csv file via the REST API, if the file contains consecutive lines without a trailing comma, the first of these lines is imported but subsequent lines without a trailing comma are not.

Bug filed as:
http://bugzilla.zimbra.com/show_bug.cgi?id=9282