[OpenMMS] Is anyone using Db Loader?

[Dan Bolser]

I found "Db Loader" - http://sw-tools.rcsb.org/ - was recently updated
(February 8, 2008). Is anyone using this software to create a
relational version of the PDB? Previously I had problems using this
tool (version 3.001). The a recent version (4.013) worked quite well
(given a recent version of MySQL).

Any info on how this compares to OpenMMS? Is there a forum for
questions about DB-Loader?


My question is about where to get a 'clean' copy of the
schema_mapping.cif file (the file that maps the mmCIF object model
into the relational database). The distribution comes with one such
mapping, hidden away under a test directory
(db-loader-v4.0/test/schema_map_pdbx_na.cif). Using this file I can
create a 'revised' mapping while processing the data files for
loading. However, I am not 100% sure what to do with the revised file.
So far I have simply switched to using the revised file to create the
MySQL schema before loading the data. However, the README suggests
that I should "update the current schema mapping configuration file"
using the "-update" command line option. I am not sure how to do this,
or if it is necessary.

I would like to be using the "correct" or "standard" schema mapping file...

Hmm... I just found 'Example 5' in the docs;


Example 5: In this example the current schema and the revised schema
mapping are combined to create a new schema mapping.

db-loader -map schema_mapping.cif -update updated_schema_mapping.cif \
          -revise revised_schema_mapping.cif

After executing this command, an updated schema mapping is stored in
the file "updated_schema_mapping.cif" and can be used in future data
processing with "-map" option.


However, after some testing I found that the MySQL schema that is
created using revised_schema_mapping.cif is no different from the one
created using updated_schema_mapping.cif (using the above naming
convention). The mapping files themselves are not identical, but the
resulting SQL schema is identical. Seems like I was (somehow) doing
things OK by first processing the data for loading (and generating a
new mapping on the way) and then generating the SQL from the 'revised'
mapping.


Cheers,

Dan.