calibredb exception

[ZioNemo]

Hi,
I'm writing a tool to automate lib management & transfer between different databases (BookCAT, in my case).
As directed I'm using calibredb to interact with Calibre database.
Basic creation of books in Calibre works
To handle updating I need to query what books are already there and their ID.
To have a better matching I wanted to use optional fields (e.g.: ISBN), but it seems calibredb chokes when one requested field is empty.

$ calibredb list  --separator='|' --fields=title,authors,isbn
id|title                   |authors      |isbn|
Traceback (most recent call last):
  File "site.py", line 58, in main
  File "site-packages/calibre/library/cli.py", line 1475, in main
  File "site-packages/calibre/library/cli.py", line 221, in command_list
  File "site-packages/calibre/library/cli.py", line 138, in do_list
  File "textwrap.py", line 329, in wrap
  File "textwrap.py", line 258, in _wrap_chunks
ValueError: invalid width 0 (must be > 0)
$ calibredb list  --separator='|' --fields=title,authors
id|title                    |authors      |
1|Calibre Quick Start Guide|John Schember|

Is this intentional? If so: how can I work around this?
I'm currently using "calibredb (calibre 1.20)" on Debian Linux (wheezy).
I know it's not the latest, but it should be recent enough, isn't it?

TiA

[kovidgoyal]

If you want machine parseable output use calibredb catalog not calibredb list. The latter is meant for useage by humans.

calibredb catalog t.csv -h

That error happens not when one of the requested fields is empty, but when it si empty for *all* the requested books.

[ZioNemo]

Nice, Thanks.
I have been using "calibredb catalog /tmp/catalog.xml --with-library=/..." which, if i'm right, dumps all non-null fields.
Is there some way to use the same format to input data?
It would be very handy to dump the catalog, modify/add to it and then reload it.
IMHO "calibredb set-metadata" format isn't as strightforward.

[kovidgoyal]

No, the only way is to scriipt it via set_metadata.