- Compatible with Pilosa 1.4
- Added additional GroupBy support.
- Fixed a bug in time view handling.
- Compatible with Pilosa 1.3
- Added support for batch key translation.
- Compatible with Pilosa 1.3
- Security Notice: A vulnerability was found in Jackson Databind 2.9.9 which was used in older versions of this library. We strongly suggest you to update to this version.
- Breaking Change: Roaring imports require nanosecond timestamps, instead of seconds.
- Compatible with Pilosa 1.3
- Fixed an issue with roaring bitmap implementation which would cause a non-optimal container type to be chosen: #5.
- Compatible with Pilosa 1.3
- Added support for Open Tracing.
- Added support for automatically loading shard width per index.
- Added
Schema.hasIndex
andIndex.hasField
methods.
- Compatible with Pilosa 1.2 and 1.3
- Added
manualServerAddress
client option which forces the client use only the manual server address. - Removed experimental
PilosaClient.withAddress
variant that sets the coordinator address manually. UsemanualServerAddress
client option instead.
- Added
field.rows
andindex.groupBy
calls. - Added
PqlBatchQuery.size
andPqlBatchQuery.clear
methods. - Deprecated
field.range
(time) call. Usefield.row
instead. - Experimental:
PilosaClient.withAddress
supports setting the coordinator address manually. - Experimental: Added
Schema.index
andIndex.field
methods support passing index/field options in a map.
- Compatible with Pilosa 1.2
- Supports imports involving keys.
- Added support for mutex and bool fields.
- Added
index.Options
,field.ClearRow
andfield.Store
functions to support the corresponding PQL calls. - Added
com.pilosa.client.csv
package. - Added support for roaring importing
RowIDColumnID
with timestamp data. - Updated
com.pilosa.roaring
dependency for improved memory usage. - Improved import speed.
- Fixed schema synchronization.
- Enabled creating the uber jar which contains the Pilosa client and all its dependencies in a single jar.
- Run
mvn -f com.pilosa.client/pom.xml -Dmaven.test.skip=true clean package
*pilosa-client-*-uber.jar
is created incom.pilosa.client/target
directory.
- Run
- Deprecated:
indexOptions.keys
, useindexOptions.setKeys
instead. - Deprecated:
indexOptions.trackExistence
, useindexOptions.setTrackExistence
instead. - Deprecated:
fieldOptions.keys
, usefieldOptions.setKeys
instead.
- Added
trackExistence
index option. - Added
not
index method to supportNot
queries. The corresponding index must be created withtrackExistence=true
option. This feature requires Pilosa on master branch. - Added support for roaring imports which can speed up the import process by %30 for non-key column imports. Pass
setRoaring(true)`` to
ImportOptions.builder()` to enable it. This feature requires Pilosa on master branch. - Fixes:
Column.create
method. See: #127
- Compatible with Pilosa 1.0.
- Following terminology was changed:
- frame to field
- bitmap to row
- bit to column
- slice to shard
- There are three types of fields:
- Set fields to store boolean values (default)
- Integer fields to store an integer in the given range.
- Time fields which can store timestamps.
- Added
keys
field option. - Experimental: Import strategies are experimental and may be removed in later versions.
- Removed all deprecated code.
- Removed
Field
type and renamedFrame
toField
.
- Compatible with Pilosa 0.9.
- Supports multi-threaded imports and import progress tracking.
- Added
RangeField.min
andRangeField.max
methods. - Deprecation
inverseEnabled
frame option,Frame.inverseBitmap
,Frame.inverseTopN
,Frame.inverseRange
methods. Inverse frames will be removed on Pilosa 1.0.
- Compatible with Pilosa master, not compatible with Pilosa 0.8.x releases.
- Checks the server version for Pilosa server compatibility. You can call
clientOptions.setSkipVersionCheck()
to disable that.
- Added
equals
,notEquals
andnotNull
field operations. - Removal
TimeQuantum
forIndexOptions
. UseTimeQuantum
of individualFrameOptions
instead. - Removal
IndexOptions
class is deprecated and will be removed in the future. - Removal
schema.Index(name, indexOptions)
method. - Removal column labels and row labels.
- Added IPv6 support.
- Added support for creating range encoded frames.
- Added
Xor
call. - Added range field operations.
- Added support for excluding bits or attributes from bitmap calls. In order to exclude bits, call
setExcludeBits(true)
in yourQueryOptions.Builder
. In order to exclude attributes, callsetExcludeAttributes(true)
. - Customizable CSV time stamp format.
- `HTTPS connections are supported.
- Deprecation Row and column labels are deprecated, and will be removed in a future release of this library. Do not use
IndexOptions.Builder.setColumnLabel
andFrameOptions.Builder.setRowLabel
methods for new code. See: FeatureBaseDB/featurebase#752 for more info.
- Fixes
filters
parameter of theTopN
parameter. - Fixes reading schemas with no indexes.
- Failover for connection errors.
- More logging.
- Uses slf4j instead of log4j for logging.
- Introduced schemas. No need to re-define already existing indexes and frames.
- make commands are supported on Windows.
-
- Breaking Change: Removed
timeQuantum
query option.
- Breaking Change: Removed
- Deprecation
Index.withName
constructor. Useschema.index
instead. - Deprecation
client.createIndex
,client.createFrame
,client.ensureIndex
,client.ensureFrame
. Use schemas andclient.syncSchema
instead.
- Supports Pilosa Server v0.4.0.
- Breaking Change: Renamed
BatchQuery
toPqlBatchQuery
. - Updated the accepted values for index, frame names and labels to match with the Pilosa server.
Union
queries accept 0 or more arguments.Intersect
andDifference
queries accept 1 or more arguments.- Added
inverse TopN
andinverse Range
calls. - Inverse enabled status of frames is not checked on the client side.
- Available on Maven Repository.
- Initial version
- Supports Pilosa Server v0.3.1.