Changelog

2.0.1

Updating to a pypi version of singer-python (realit-singer-python)

2.0.0

Updating to a patched version of pipelinewise-singer-python using msgspec instead of orjson for serialization.
Speeding up tap-oracle via the use of new pipelinewise-singer-python library.
deprecating support for Python 3.7, and adding support for Python 3.11 and 3.12 via pipelinewise-singer-python.

1.3.0

Adding new optional configuration setting 'ora_python_driver_type'. This setting allows you to pick which Oracle Library Driver / mode to operate in.
cx : cx_Oracle (Use the legacy cx_Oracle library - default)
thin: oracledb (Use thin mode - no Oracle Client required. Required for MacOS)
thick: oracledb (Use thick mode - use Oracle Client)

1.2.5

Resolving bug in the filter_sys_or_not function to handle an empty schema filter. The fix also allows the SYS schema to be anywhere in the list.

1.2.4

Breaking Change! Removing logic to assuming an Oracle number(1) datatype is a boolean.

1.2.3

Add tags to the record_count (database / schema)
Adding logo for pipelinewise-tap-oracle
Fix make sure key_properties is not null for views

1.2.2

Handling NULL's / None in Singer.Decimal columns.

1.2.1

Applying cursor array_size to incremental and log_based.
Renaming config parameter from full_table_sync_batch_size to cursor_array_size.
Increasing the default array size from 100 to 1000.

1.2.0

New config option to provide an offset for incremental loads - offset_value.
Changing the sort order to sort by the column_id so tables columns match the database source.

1.1.9

Pulling the database name from the env if v$database is unavailable.

1.1.8

Swapping singer-python for pipelinewise-singer-python
This variant uses orjson for serializing 40-50x faster than other libraries.

1.1.7

Bumping cx_Oracle to 8.3
Removing unnecessary call to get the database name

1.1.6

Added support to emit numeric data using the singer.decimal notation to avoid numeric rounding issues.
To enable singer.decimal notation add the following config "use_singer_decimal": true

1.1.5

Large change to incorporate a number of cherry-picked features plus new features, table_filtering, no ora_rowscn full table loads.
- Reverting commits for treating decimals and floats as singer.decimal.
- Bumping cx_Oracle to 8.2
- Adding query timeouts
- Adding support to connect via service_name
- Dynamically reduce the SCN_WINDOW_SIZE with timeouts occur
- Support for Plugable database connections
- Bumping singer-python to version 5.12.2
- Allow full_table with no ORA_ROWSCN and order by clause. Note: Not restartable.
- Adding Datetime, Date, NCLOB, CLOB, and BLOB datatypes
- Discovery filter to set tables via ENV MELTANO_EXTRACT__SELECT or config item filter_tables.

1.1.2

Log value of mine_sql #30

1.1.1

Set a maximum length on Singer Decimals, where a decimal past the cap is normalized via decimal.normalize() #28

1.1.0

Values with Decimal precision will now be written as strings with a custom singer.decimal format in order to maintain that precision through the pipeline #26

1.0.1

Increase default numeric scale from 6 to 38 #24

1.0.0

Backwards incompatible change to the way that data types are discovered and parsed #22
- Oracle numeric types with a null scale (NUMBER and NUMBER(*)) will now be correctly discovered as floating point types rather than integers.
- This may cause downstream issues with loading and reporting, so a major bump is required.

0.3.1

Adds handling for columns that do not have a datatype -- those columns will have inclusion=unavailable and sql-datatype="None" #19

0.3.0

Adds optional parameter scn_window_size to allow for an scn window during logminer replication #18

(2020-05-19) - Pipelinewise 1.0.1

Fixed an issue when output messages were not compatible with pipelinewise-transform-field component

(2019-09-08) - Pipelinewise 1.0.0

Initial release and fork of singer tap-oracle 0.2.0
use trap_stream_id as stream in singer messages to make it compatible with PipelineWise components