ldif-csv-conv — LDIF to CSV and vice versa

These two scripts can convert datasets between the LDIF (LDAP Data Interchange Format) and CSV (Comma-Separated Values) formats, in both directions, without losing information.
(Comments in the LDIF are not preserved; ordering of multi-value attributes is preserved.)

LDIF to CSV

./ldif2csv data.ldif > data.csv

or

./ldif2csv < data.ldif > data.csv

CSV to LDIF

./csv2ldif data.csv > data.ldif

or

./csv2ldif < data.csv > data.ldif

Notes (format, etc)

• The CSV files use ; as the cell separator, enclose all cells in " (with "s in the input represented as "") and separate the values of a multi-value attribute with |:

  "dn";"email";"objectclass" "uid=johndoe,dc=example,dc=com";"johndoe@example.com";"inetOrgPerson|top"

  The "" escaping allows you to use both ; and " in your data. If you need to use |, you'll have to edit the JOINER definition in both script files.
  ldif2csv will examine your data and output suggestions for unused characters that could be used as joiners or delimiters.

• Individual attribute values¹ longer than 499 bytes (think jpegphotos) are split off into individual files to keep the CSV small.
  They are referenced as FILE=dn=…,attr=…,id=… in the respective CSV cell. If this format is kept intact, they will be reintegrated into the LDIF by csv2ldif.
  The user is free to manually insert or modify these FILE=… references — if they point to a valid file, they will be followed; if not, the converter will crash and burn. Enjoy!
  (¹ Note that this means you can have a mixture of "direct" and "referenced" values in the same cell.)

Known issues

• Conversion from CSV to LDIF is very slow when the input contains a lot of binary data (e.g. jpegphoto attributes).
• Error handling? What's that?
  If something goes wrong, you'll get your usual Python exception stack trace.