The verbosity is not that easy to adjust. Remove useless messages and add a
new one telling when the COPY of a table is done. As we might have to wait
for some time for indexes being built. keep the CREATE INDEX lines. Also
keep the ALTER TABLE both for primary keys and foreign keys, again because
the user might have to wait for quite some time.
In the complete PostgreSQL schema step, an error would be logged as you
expect but poorly handled: it would have the whole transaction rolled back,
meaning that a single Primary Key definition failure would cancel all the
others, plus the foreign keys, and also the triggers and comments.
It happens that other systems allow a primary column to contain NULL values,
which is forbidden in the standard and enforced by PostgreSQL, so that's not
a theoritical concern here.
In the prepare-pgsql-database method we were logging too much details, such
as DDL warnings on if-not-exists for successful queries. And those logs are
to be found in PostgreSQL server logs anyway.
Also fix trying to create or drop a "nil" schema.
Given new SQLite test case from issue #563 we see that pgloader doesn't
handle errors gracefully in post-copy stage. That's because the API were not
properly defined, we should use pgsql-execute-with-timing rather than other
construct here, because it allows the "on error resume next" behavior we
want with after load DDL statements.
See #563.
When the option "drop indexes" is in use in loading data from a file, we
collect the indexes from the PostgreSQL catalogs and then issue DROP
commands against them before the load, then CREATE commands when it's
done.
The CREATE is done in parallel, and we create an lparallel kernel for
that. The kernel must have a worker-count of at least 1, and we where
not considering the case of 0 indexes on the target table.
Fix#484.
In cases where we have a WITH include drop option, we are generating
lots of SQL DROP statements. We may be running an empty target database
or in other situations where the target object of the DROP command might
not exists. Add support for that case.
Calling a -with-timing from within a with-stats-collection macro is
redundant and will have the numbers counted twice. Which in this case
didn't happen because the stats label was manually copied, but borked
with a typo in one copy.
When loading data into an existing PostgreSQL catalog, we DROP the
indexes for better performance of the data loading. Some of the indexes
are UNIQUE or even PRIMARY KEYS, and some FOREIGN KEYS might depend on
them in the PostgreSQL dependency tracking of the catalog.
We used to use the CASCADE option when dropping the indexes, which hides
a bug: if we exclude from the load tables with foreign keys pointing to
tables we target, then we would DROP those foreign keys because of the
CASCADE option, but fail to install them again at the end of the load.
To prevent that from happening, pgloader now query the PostgreSQL
pg_depend system catalog to list the “missing” foreign keys and add them
to our internal catalog representation, from which we know to DROP then
CREATE the SQL object at the proper times.
See #400 as this was an oversight in fixing this issue.
Also known as the ORM case, it happens that other tools are used to
create the target schema. In that case pgloader job is to fill in the
exiting target tables with the data from the source tables.
We still focus on load speed and pgloader will now DROP the
constraints (Primary Key, Unique, Foreign Keys) and indexes before
running the COPY statements, and re-install the schema it found in the
target database once the data load is done.
This behavior is activated when using the “create no tables” option as
in the following test-case setup:
with create no tables, include drop, truncate
Fixes#400, for which I got a test-case to play with!
