Upgrading From Earlier Versions

Catalogs generated with repoze.catalog versions prior to 0.8.0 require migration in order to add some house keeping data to the indexes that was not needed prior to 0.8.0. The easiest way to migrate a catalog is actually to delete the catalog and reindex your content objects. This is sufficient in the vast majority of cases, since the catalog merely indexes data which itself lives outside of the catalog, either in the same ZODB instance or in another database altogether.

For particularly large datasets, rebuilding your catalog from scratch can potentially take a long time. In order to make this less painful, repoze.catalog does contain a migration utility which can migrate an existing catalog in a much shorter period of time. This utility comes in three different flavors which can be used according to your use case. The primary problem being solved by the migration utility is making the indexes aware of all of the docids stored in the catalog, not just just the docids of documents which have values in a particular index. To this end, the three different flavors of migration utility represent three different ways to obtain the set of all document ids stored in the catalog.

Migrate Using a Document Map

It will be the easiest and fastest to migrate your pre-0.8.0 catalog if you are using a document map which is an instance of repoze.bfg.document.DocumentMap:

from repoze.catalog.migration import migrate_to_0_8_0_from_document_map

migrate_to_0_8_0_from_document_map(catalog, document_map)

Migrate Using Self-Generated Set of Docids

Alternatively, if you do not use a document map, you can use a self-generated set of document ids. It is up to your application to calculate the set of all document ids used in the catalog:

from repoze.catalog.migration import migrate_to_0_8_0_from_docids

migrate_to_0_8_0_from_docids(catalog, docids)

Let the Migration Utility Calculate the Set of Docids

Your last option is to allow the migration utility to calculate the set of all docids by examinging the indexes in the catalog. If you can assume that every document indexed by the catalog has a value in at least one index, then this method, while potentially slow, should succeed in calculating the superset of all document ids in the catalog:

from repoze.catalog.migration import migrate_to_0_8_0