Changes
Version 0.2.3, 2020-05-02
Fixed
ForeignKeyViolation
when building normalized temporary tablecharacters_to_character
.Fixed
ValueError
when no command was specified for the pimdb command line client.
Version 0.2.2, 2020-04-26
Fixed
AssertionError
when command line option--bulk
was less than 1.Added NAME
normalized
as option for pimdb transfer to transfer only the datasets needed by pimdb build.Removed redundant normalized tables
title_to(director|writer)
. Use relationpraticipation.profession_id
to limit query results to certain professions.Added documentation chapter explaining the Data model including example SQL queries and overview ER diagrams.
Added automatic removal of temporary tables only needed to build the normalized tables.
Version 0.2.1, 2020-04-18
Improved performance of command build for PostgreSQL by changing bulk
insert
tocopy from
.
Version 0.2.0, 2020-04-16
Fixed command build for PostgreSQL (issue #25).:
Index names now have at most 63 characters under PostgreSQL. Proper limits should also be in place for MS SQL and Oracle but have yet to be tested. SQLite always worked because it has a very large limit.
The PostgreSQL docker container for the test run now has more shared memory in order to allow “insert … from select …” with millions of rows. Performance still has a lot of room for improvement.
Added TV episodes (tables
TitleEpisode
resp.episode
).Cleaned up logging for
transfer
andbuild
to consistently log the time and rows per second for each table.
Version 0.1.2, 2020-04-14
Fixed remaining “value to long” errors (issue #14).
Fixed
TypeError
when command line option –bulk was specified.Added instructions on how to test run pimdb on a PostgreSQL docker container, see Test run with PostgreSQL.
Version 0.1.1, 2020-04-13
Fixed “value to long” for
NameBasics.knowForTitles
(issue #13).Added option to omit “sqlite:///” prefix from
--database
and specify only the path to the database file.Moved documentation to ReadTheDocs.
Improved performance of SQL inserts by using bulk inserts consistently and changing loops to SQL
insert ... from select ...
(where possible).
Version 0.1.0, 2020-04-11
Initial public release.