• maint

• add -safe flag following #5

Maintenance release.

• 1.6.2 release
• use batched operations in marctojson
• TODO: eventually apply optimizations to marcxmltojson and marctotsv

Added marcsnapshot command.

Status: Experimental.

marcsnapshot takes a list of MARC files and deduplicates records, keeping only the "newest", if multiple versions are found.

Field 001 is used as ID. Field 005 value is used to sort.

Example: three marcfiles, possibly with duplicates.
Field 001 values can be identified by the first 9 chars.

    $ marcsnapshot -verbose -l 9 2010.mrc 2011.mrc 2012.mrc > out.mrc

The default output of marctojson uses "record" as key, not "content". If you rely on the key being "content", use

$ marctojson -recordkey content ...

to get the old behaviour.

• marctojson got a -recordkey flag
• move marctools import from miku to ubleipzig

marcdb got an -encode flag.

Small improvements in marcdb.

Added marcdb command.

Added marcxmltojson.

Changed json export format for marctojson.

The new format uses arrays for all subfield values,
not just for those, which had more than one element.

old: {"020": [{"a": "1234567890", "ind1": " ", "ind2": " ", "z": ["A", "B"]}]}
new: {"020": [{"a": ["1234567890"], "ind1": " ", "ind2": " ", "z": ["A", "B"]}]}

This unifies and hopefully simplifies processing on the consumer side.

Control field values are still just strings, since they are single values by spec.