- Remove
maxLength
from the catalog #88
- Removes rowversion and replication keys as valid interruptible bookmarks for logical replication, ensuring the use of a Primary Key
- Adds additional check for valid Primary Key before beginning logical replication rather than after completing initial historical replication
- Updates full table interruptible bookmark key priority to optimize index efficiency by using PK if available before falling back to rowversion if available
- Adds additional tests for full-table interruptible
- #86
- Update singer-clojure and logging dependencies #84
- Fix RowVersion states for logical syncs #76
- Update
nrepl
server binding from0.0.0.0
to127.0.0.1
#62
- Quote the
ORDER BY
columns in incremental queries #59
- Attempt readonly connections on
jdbc/reducible-query
andjdbc/query
calls in Full Table and Incremental Syncs #58
- Apply approaches from 1.6.5 and 1.6.6 to only try
ApplicationIntent=ReadOnly
for query-based connections, and fall-back to not read only if the check fails. #55
- Handle cases where
ApplicationIntent=ReadOnly
is not doable for log-based sync's initial full table #53
- Only adds
ApplicationIntent=ReadOnly
to query-based connections due to an issue with Change Tracking and secondary read replicas not supporting it #52
- Adds
ApplicationIntent=ReadOnly
to the connection string #50
- Properly check whether change tracking is enabled on a DB in Azure #41
- Warn on permissions errors when discovering schemas if the user doesn't have access #33
- Fix bugs with Views being interrupted during a full table sync #28
- Configure ResultSet options for concurrency mode and cursor type #25
- Clarify error messages in edge case when change tracking is not available #24
- Fix edge case where two identically named tables in different schemas have different change tracking status #24
- Fix issue where some datetime types were having issues when approaching year 0. #22
- During query generation surround schema names with square brackets to allow reserved words in schemas. #20
- Added a fallback value for
_sdc_deleted_at
when running a log based sync. - The tap also logs when this happens.
- #18
- Fixes a bug where during discovery, for columns of type
binary
the tap was writing the schema as a string, but not transforming the data to a string, instead emitting it as a byte array. #16
- Fix a bug where timestamp column bookmarks cause an exception when resuming full-table #15
- Extract database from config replacing it with an empty string if it is nil #13
- Removed max length from binary-type JSON schemas #8
- Add support for connecting to Named Instances by omitting port and adding the instance name in the host field #4
- Fix discovery query to correctly type
sys.partitions.rows
asbigint
#5
- Add support for money and smallmoney columns #1
- Fix the sql query generation for full table interruptible syncs with composite pks #53
- Fix a bug with a View's key_properties being improperly set #51
- Add an assertion to ensure log-based replication has the primary keys needed to replicate #50
- Make _sdc_deleted_at nullable commit
- Add support for datetime2, datetimeoffset, and smalldatetime #49
- Add support for numeric and decimal identity columns #48
- GA Release