This package contains tools for performing batch imports to the Vertica database.
pyvertica.connection
- Module which contains logic for connection to a Vertica DB.
pyvertica.batch
- High speed loader for Vertica.
pyvertica.importer
- Base-class for writing Vertica batch importers.
pyvertica.migrate
- Module managing export from one Vertica cluster to another.
vertica_batch_import
- Command-line interface for the
pyvertica.batch
module. vertica_migrate
- Command-line interface for the
pyvertica.migrate
module.
pyvertica can be installed by executing pip install pyvertica
.
- UPDATE
TaskThread
requred version to 1.3 or greater.
- FEATURE
VerticaBatch
will not open a new thread or fifo each time a new batch is started ifmulti_batch
is set toTrue
. This will reduce the overhead each time the batch is committed, provided the user wants to callcommit
more than once on a single batch.
- ADD:
insert_lists
method toVerticaBatch
. This method takes multiple lists, converts them to rows, and invokes a single insert rather than one insert per row.
- ADD:
connection
paratmeter toVerticaBatch
to allow usage of an existing connection.
- Use latest pyodbc version.
- ADD:
rolllback
method toVerticaBatch
object. - ADD:
REJECTMAX
option. - CHANGE:
VerticaBatch.get_errors
now returns the number of errors instead of abool
indicating if there are errors. - CHANGE:
REJECTEDFILE
option is now set to__debug__
.
- CHANGE: Make it possible to pass host, and credentials to the
vertica_migrate
tool by using an config file (to not expose credentials on the command-line).
- CHANGE: Make it more easy to pass more arguments to
get_connection
throughBaseImporter
andVerticaBatch.
Note that this is a backwards incompatible change as the arguments forBaseImporter
,VerticaBatch
andget_connection
have changed.
- FEATURE: Add a migration module
pyvertica.migrate``and script ``vertica_migrate
, to move data between clusters. - CHANGE: Make the reconnect optional in
VerticaBatch.get_connection
- CHANGE: Make the execution of
ANALYZE_CONSTRAINTS
optional (executed when getting the errors).
- FIX: Handle exceptions raised inside the
QueryThread
so that the semaphore always gets released.
- FIX: Line-endings in file-object returned by
get_errors
. All is now\n
.
- CHANGE: The Batch-history table is now configurable in
BaseImporter
. - CHANGE:
get_connection
selects a random node from the cluster and returns a connection to that node. - FIX: Cleanup fifo + temporary directory.
- CHANGE: Change the way in how we detect if the FIFO object is consumed
by the QueryThread. By opening the FIFO object (for writing) in
'w'
mode after the QueryThread was created, it will block until the COPY SQL statement is started.
- Public release!