Skip to content
Find file
598 lines (558 sloc) 32.8 KB

Scalding

Version 0.16.0

  • Add tests around hashcode collisions : #1299
  • Fix performance bug in TypedPipeDiff : #1300
  • make serialization modules build on travis : #1301
  • Improve TypedParquetTuple : #1303
  • Add UnitOrderedSerialization : #1304
  • Revert "Add UnitOrderedSerialization" : #1306
  • Change groupRandomly & groupAll to use OrderedSerialization : #1307
  • Make test of Kmeans very very unlikely to fail : #1310
  • make LongThrift sources TypedSink : #1313
  • Fix testing VersionedKeyValSource#toIterator for non-Array[Byte] types : #1314
  • Make SketchJoin ordered serialization aware : #1316
  • Added a sealed trait ordered serializer. When it works its great. Not as reliable as we'd like. But hopefully restrictions on it will do the job : #1320
  • Add secondary sorting using ordered serialization : #1321
  • Bails out from the length calculation if we don't succeed often : #1322
  • increased number of box instances to 250 : #1323
  • Apply merge strategy for pom.properties files : #1325
  • Apply merge strategy for pom.xml files : #1327
  • Add a OrderedSerialization.viaTransform with no dependencies, and a BijectedOrderedSerialization in scalding core : #1329
  • Precompute int hashes : #1330
  • Hide the deprecated string error for getting ASCII bytes. : #1332
  • Change defaults for Scalding reducer estimator : #1333
  • Execution id code : #1334
  • Add line numbers at .group and .toPipe boundaries : #1335
  • Ordered Serialization macros for thrift : #1338
  • make some repl components extensible : #1342
  • Remove the bootstrap section : #1346
  • Fix the execution test : #1347
  • Implement flatMapValues method : #1348
  • Consistent style in homepage example : #1349
  • Serialization folding : #1351
  • Collapses scalding-db packages : #1353
  • Merge scalding-macros into scalding-core : #1355
  • Migrate typedtext : #1356
  • Runtime reducer estimator : #1358
  • Update Build.scala : #1361
  • Allow overriding of hadoop configuration options for a single source/sink : #1362
  • Missing an extends Serializable, causes issues if capture Config's anywhere : #1365
  • Fix TypedPipe.limit to be correct, if slighly slower : #1366
  • Fix scala.Function2 showing up in line numbers : #1367
  • Drop with MacroGenerated from Fields macros : #1370
  • Fix deprecation warnings in TypedDelimited : #1371
  • Ianoc/revert changes around making file systems : #1372
  • Revert typed tsv behavior : #1373
  • A serialization error we were seeing in repl usage : #1376
  • Add NullSink and test : #1378
  • Add monoid and semigroup for Execution : #1379
  • Upgrade parquet to 1.8.1 : #1380
  • Upgrade sbt launcher script (sbt-extras) : #1381
  • Just move whitespace, add comments, simplify a few methods : #1383
  • Don't publish maple when doing 2.11 so we only publish it once -- needed for cross publishing to maven repo's : #1386
  • Support nesting Options in TypeDescriptor : #1387
  • Enable Scalding-REPL for Scala 2.11 : #1388
  • Updates for some upstream fixes/changes : #1390
  • Remove use of hadoop version in estimators : #1391
  • Set hadoop version to dummy value : #1392
  • Handle no history case in RatioBasedEstimator : #1393
  • Inline parquet-scrooge : #1395
  • RatioBasedEstimator - fix threshold edge case, add tests : #1397
  • Fixes the scrooge generator tasks not to generate code in the compile target, we were publishing these : #1399
  • Ianoc/configure set converter : #1400
  • Change hash function in GroupRandomly : #1401
  • Improve logging in runtime reducer estimators : #1402
  • Add the type in ScroogeReadSupport : #1403
  • Adds a function to test if a sink exists at the version we created : #1404
  • add .groupWith method to TypedPipe : #1406
  • Add some return types : #1407
  • add counter verification logic : #1409
  • Runtime reducer estimator fixes : #1411
  • Make sure Execution.zip fails fast : #1412
  • When using WriteExecution and forceToDisk we can share the same flow def closer in construction : #1414
  • Cache the zipped up write executions : #1415
  • Fix DateOps "match may not be exhaustive" warning : #1416
  • Factor out repeated code into FutureCache : #1417
  • Fix lack of Externalizer in joins. : #1421
  • Adds much more line number information through the NoStackAndThen class : #1423
  • Requires a DateRange's "end" to be after its "start" : #1425
  • Scalding viz options : #1426
  • Fixes map-only jobs to accommodate both an lzo source and sink binary converter : #1431
  • Fix Readme travis link : #1432
  • Fixes docs wording : #1433
  • Don't squash the exception in history service when there's a failure : #1434
  • Log the exception in RatioBasedEstimator when there's a failure : #1435
  • make getBytesPerReducer support human readable values like 128m and 1g : #1436
  • Fixes minor KeyedList docs wording : #1437
  • Fix readPathsFor to use the tz argument : #1439
  • Scalding viz options : #1440
  • call Job.validate when running tests under JobTest : #1441
  • opt-in to calling Job.validate in JobTest : #1444
  • Fix bug with sketch joins and single keys : #1451
  • Fix incorrect usage of percent. : #1455
  • Add OrderedSerialization2 support in Matrix2. : #1457
  • Add InvalidSourceTap to catch all cases for no good path. : #1458
  • Cluster info and fs shell in repl : #1462
  • Update Scala version to 2.10.6 : #1463
  • Fix median estimation : #1464
  • Makes the config transient in the KryoHadoop instanciator : #1466
  • Moves the default to 2.11 : #1467
  • Adds Error Message to REPL when Current Directory Not Readable : #1468
  • SuccessFileSource: correctness for multi-dir globs : #1470
  • Limit task history fields consumed from hraven : #1472
  • Remove dependency on dfs-datastores : #1473
  • ScaldingILoop should enable one to pass in in/out : #1475
  • Switch Chat to Gitter : #1477
  • Add two functions that assist in testing a TypedPipe : #1478
  • Makes permission failures non-fatal when looking for .scalding_repl files : #1479
  • Update TypeDescriptor to explain that Option[String] is not supported : #1480
  • Remove a type parameter that doesn't seem to do anything : #1481
  • Utility for expanding libjars : #1483
  • Shouldn't skip hidden files, user can decide such things with their glob : #1485
  • Fix FileSystem.get issue : #1487
  • Remove dependency on parquet-cascading : #1488
  • Add withConfig api to allow running an execution with a transformed config : #1489
  • Call validateTaps in toIterator codepath : #1490
  • Update the build : #1491
  • Arg Descriptions/Help for Execution Apps : #1492
  • Fix issue #1429 : #1493
  • Cache counters for stat updates : #1495
  • Pulls the core ExecutionTests back into scalding-core : #1498
  • Add a liftToTry function to Execution : #1499
  • Small improvements to the Boxed.scala module : #1500
  • Cache boxed classes : #1501
  • Fix unnecessary use of .get in Globifier.scala : #1502
  • Replace unintentional use of Unit with () : #1503
  • Fix unnecessary uses of Option.get : #1506
  • Utility methods for running Executions in parallel : #1507
  • Typed Mapside Reduce : #1508
  • Use wartremover to guard against careless use of _.get : #1509
  • Add in an API around cache isolation : #1511
  • Add implicit Ordering[RichDate] : #1512
  • Fix MultipleTextLineFiles source in JobTest : #1513
  • Add's support for sealed abstract classes : #1518
  • Update FixedPathSource to strip out '' in paths ending with '/' for writes : #1520
  • support for more formats to work with RichDate : #1522
  • WIP: Add forceToDisk parameter to hashJoin in TypedPipe : #1529
  • Fixing comments on partitioned delimited source : #1530
  • Remove weakly typed Source : #1531
  • Maple fix for HBaseTap : #1532
  • Add an enrichment for TypedPipe.inMemoryToList and use it in TypedPipeDiff test : #1533
  • Because, because... fun, the scala compiler has special naming rules it appears when there are leading underscores : #1534
  • Fix README examples link : #1536
  • Fixes Config to accommodate spaces in argument values : #1537
  • Add before() and after() to RichDate : #1538
  • Adds late tap validation for cases where race conditions cause it to fail : #1540
  • Fix Rounding Bug in RatioBasedEstimator : #1542

Version 0.15.0

  • Move OrderedSerialization into zero-dep scalding-serialization module #1289
  • bump elephantbird to 4.8 #1292
  • Fix OrderedSerialization for some forked graphs #1293
  • Add serialization modules to aggregate list #1298

Version 0.14.0

  • add .unit to Execution object #1189
  • Override hashCode for Args #1190
  • Put a value in a exception message #1191
  • Add an exclusiveUpper method to DateRange #1194
  • Covert LzoTextDelimited to Cascading scheme. #1179
  • Remove Travis IRC notifications #1200
  • add LookupJoin and LookupJoinTest changes from summingbird #1199
  • Add a new ExecutionApp tutorial #1196
  • Move main simple example to be the typed API, and put the .'s at the sta... #1193
  • Add Execution.withArgs #1205
  • Config/Cascading updater #1197
  • Remove algebird serializers #1206
  • remove warnings in CumulativeSum #1215
  • Implicit execution context / easier switching between modes #1113
  • add row l1 normalize #1214
  • provide Args as an implicit val #1219
  • call sourceConfInit when reading from taps in local mode #1228
  • Add distinctCount and distinctValues helper methods to KeyedList. #1232
  • import hygiene: remove unused imports and remove JavaConversions use #1239
  • Swap hash and filename for filename-extension-sensitive code #1243
  • Remove more unused imports #1240
  • Provide useHdfsLocalMode for an easy switch to mapreduce local mode #1244
  • upgrade scalacheck and scalatest #1246
  • Optimize string and (hopefully) number comparisons a bit #1241
  • Note the active FlowProcess for Joiners #1235
  • Make sure Executions are executed at most once #1253
  • Fix Config.getUniqueIDs #1254
  • Add MustHasReducers trait. #1252
  • Make sure the EvalCache thread isDaemon #1255
  • Use non-regex split function #1251
  • make InputSizeReducerEstimator work for any CompositeTap #1256
  • TimePathedSource helper methods #1257
  • Fix for reducer estimation not working correctly if withReducers is set to 1 reducer #1263
  • Add make(dest) to TypedPipe #1217
  • Fix SimpleDateFormat caching by default #1265
  • upgrade sbt and sbt launcher script #1270
  • Add TypedPipeDiff for comparing typed pipes #1266
  • Change separator from \1 to \u0001 #1271
  • Disable reducer estimation for map-only steps #1276
  • Local sources support multiple paths #1275
  • fix the spelling of the cumulativeSumTest file #1281
  • Hydrate both sides of sampledCounts in skewJoinWithSmaller #1278
  • Bijection 0.8.0, algebird 0.10.0, chill 0.6.0, scala 2.10.5 #1287
  • Remove some deprecated items #1288

Version 0.13.1

Version 0.13.0

Version 0.12.0

Version 0.11.2

  • hadoop.tmp.dir for snapshot in config

Version 0.11.1

Version 0.11.0

Version 0.10.0

  • Upgrade cascading to 2.5.4, cascading jdbc to 2.5.2
  • Adding an hdfs mode for the Scalding REPL
  • Added implementation of PartitionSource with tests
  • Add helper methods to KeyedList and TypedPipe
  • Add addTrap to TypedPipe

Version 0.9.0

  • Add join operations to TypedPipe that do not require grouping beforehand
  • Fixed bug in size estimation of diagonal matrices
  • Optimized the reduceRow/ColVectors function for the number of reducers
  • Add a BlockMatrix object (an abstraction of a Vector of Matrices)

Version 0.8.8

  • Publish 0.8.7 for scala 2.9.3.

Version 0.8.7

  • Hotfix to bypass a bug in Hadoop, which cannot sync up all deprecated keys.

Version 0.8.6

  • Hotfix a bug in Tool. Now Tool will re-throw all exceptions again.

Version 0.8.5

  • Fixed bug in RichPipe.insert
  • Add souce[T] to JobTest
  • Allow DelimitedScheme to override strictness and safety.
  • Add distinct method to RichPipe and TypedPipe, add mapValues to TypedPipe
  • ISSUE 389: Catch exceptions in Tool
  • Sbt assembly 0.8.7
  • Add CascadeJob to allow multiple flows in one job
  • Adding cross-build scala versions
  • Use mima to check binary compatibility

Version 0.8.4

  • ISSUE 340: Upgrade to Cascading 2.1.5
  • ISSUE 327,329,337: adds sample method with seed in RichPipe
  • ISSUE 323: Remove untyped write from TypedPipe (must write to Mappable[U])
  • ISSUE 321: pulls out scalding-date and scalding-args as separate projects

Contributions

27 commits

  • P. Oscar Boykin: 13 commits
  • willf: 4 commits
  • Argyris Zymnis: 3 commits
  • Sam Ritchie: 3 commits
  • Tim Chklovski: 1 commits
  • Chris Severs: 1 commits
  • Rickey Visinski: 1 commits
  • David Shimon: 1 commits

Version 0.8.3

  • ISSUE 312: dramatic speedup for sortWithTake/sortedTake if you take many items
  • ISSUE 307: Read support for JsonLine support (previously, just write)
  • ISSUE 305: Adds a shuffle-method to RichPipe (for sampling/sharding)
  • ISSUE 299: limit method for Typed-safe API.
  • ISSUE 296: Fixes self-joins in the Type-safe API.
  • ISSUE 295: unpack-all syntax (use Fields.ALL) for TupleUnpacker
  • ISSUE 280: Improvements to AbsoluteDuration
  • ISSUE 277: Upgrade to Cascading 2.0.7

Contributions

75 commits total.

  • P. Oscar Boykin: 25 commits
  • Alex Dean: 15 commits
  • Argyris Zymnis: 11 commits
  • Sam Ritchie: 4 commits
  • Timothy Chklovski: 4 commits
  • Dan McKinley: 4 commits
  • Aaron Siegel: 3 commits
  • Tim Chklovski: 3 commits
  • Ashutosh Singhal: 2 commits
  • João Oliveirinha: 2 commits
  • Arkajit Dey: 1 commits
  • Avi Bryant: 1 commits

Version 0.8.2

  • ISSUE 269: Improvements to AbsoluteDuration.fromMillisecs and some new APIs
  • ISSUE 256: Weighted page-rank with the Matrix API
  • ISSUE 249: Fix for Matrix.scala missing some obvious operations
  • ISSUE 246: Partition in RichPipe (create a new field, and then groupBy on it)
  • ISSUE 241: Fix joinWithLarger with a custom joiner
  • ISSUE 234. 238: Etsy sync: periodic date jobs, ability to add traps, more flexible Args
  • ISSUE 230: shard and groupRandomly on RichPipe
  • ISSUE 229: Initial skew-Join implementation (please test!)
  • ISSUE 228 - 233, 265: Improve Typed-API
  • ISSUE 221: Combinatorics in scalding.mathematics

Contributions

106 commits total.

  • P. Oscar Boykin: 29 commits
  • Krishnan Raman: 16 commits
  • Arkajit Dey: 15 commits
  • Avi Bryant: 7 commits
  • Edwin Chen: 6 commits
  • Aaron Siegel: 5 commits
  • Koert Kuipers: 5 commits
  • Argyris Zymnis: 4 commits
  • Sam Ritchie: 4 commits
  • Chris Severs: 4 commits
  • Brad Greenlee: 3 commits
  • Wil Stuckey: 2 commits
  • Dan McKinley: 2 commits
  • Matteus Klich: 2 commits
  • Josh Devins: 1 commits
  • Steve Mardenfeld: 1 commits

Version 0.8.1

  • ISSUE 220: Etsy date improvements and local-mode tap improvements
  • ISSUE 219: scald.rb fix
  • ISSUE 218: Add aggregate method to ReduceOperations
  • ISSUE 216: Improve variance notations
  • ISSUE 213,215: Make Field[T] serializable
  • ISSUE 210,211: Refactor date code into individual files + tests
  • ISSUE 209: Add hourly/daily time pathed source classes
  • ISSUE 207,208: sbt build improvements
  • ISSUE 205,206: Remove scala serialization code to com.twitter.chill
  • ISSUE 203: Improved date-parsing and docs (from Etsy)
  • ISSUE 202,204: Add propagate/mapWithIndex in Matrix (use Monoids with graphs)
  • ISSUE 201: add stdDev to groupBuilder.
  • ISSUE 200: typed write and key-value swap
  • ISSUE 196: Clean up deprecations
  • ISSUE 194,195: Fix negative numbers as args
  • ISSUE 190-192,197,199: Improved Joining/Co-grouping in the typed-API

Version 0.8.0

  • Many small bugfixes
  • ISSUE 189: Adds spillThreshold to GroupBuilder to tune memory usage
  • ISSUE 187: Adds TypedTsv for type-safe TSV files.
  • ISSUE 179: Add forceToDisk to help hand optimization of flows
  • ISSUE 175: API to set the type/comparators in the Fields API.
  • ISSUE 162, 163: Adds keys, values methods to Typed-API
  • ISSUE 160: adds approxUniques to GroupBuilder
  • ISSUE 153: mapPlusMap in GroupBuilder
  • ISSUE 149: Support for Hadoop sequence files
  • ISSUE 148: Matrix API
  • ISSUE 147: Move Monoid/Algebra code to Algebird
  • Mulitple issues: many new Kryo serializers added
  • ISSUE 140: Adds ability to do side-effects (foreach, using in RichPipe)

Version 0.7.3

  • ISSUE 134: Cleans up scalding.Tool
  • ISSUE 133: Adds SortedListTake monoid, Either monoid, and tests
  • ISSUE 130: Upgrades cascading.kryo to 0.4.4
  • ISSUE 129: Disable Kryo references by default, add config
  • ISSUE 128: Minor fix to normalize job

Version 0.7.2

  • ISSUE 127: Upgrade maple and remove jdks.
  • ISSUE 126: Switch normalize to use crossWithTiny
  • ISSUE 125: Serialization fixes

Version 0.7.1

  • ISSUE 124: Add crossWithSmaller and use it in normalize.
  • ISSUE 123: Adds generated classes that allow tuple concatenation.

Version 0.7.0

  • ISSUE 122: Upgrade cascading to version 2.0.2 and maple to 0.2.1
  • ISSUE 119: Iterator fix
  • ISSUE 118: Specialized tuplegetter
  • ISSUE 117: Job conf keys
  • ISSUE 116: Feature/flatten
  • ISSUE 115: Fold scan init fix
  • ISSUE 114: Feature/scald update
  • ISSUE 113: Make sure SequenceFile uses the fields list if one is passed in
  • ISSUE 110: Upgrade Kryo to 2.16.
  • ISSUE 107: Add a method to GroupBuilder to force all aggregation on the reducers.
  • ISSUE 105: Feature/head fix
  • ISSUE 104: Feature/default date improvements
  • ISSUE 103: Feature/case case pack
  • ISSUE 100: Adds trivial UnitGroup
  • ISSUE 99: Fix build breakage and add tutorial run script.

Version 0.6.0

  • ISSUE 98: Feature/tpipe stream
  • ISSUE 97: Add the user input to the unknown mode error message
  • ISSUE 95: Feature/buffer cleanup
  • ISSUE 94: Feature/tpipe
  • ISSUE 93: Upgrade to cascading 2.0.0 and maple 0.2.0
  • ISSUE 92: Feature/dsl refactor

Version 0.5.4

  • ISSUE 91: Upgrade to cascading wip-310 and maple 0.1.10
  • ISSUE 90: Test against multiple JDKs on travis-ci.org
  • ISSUE 88: Bump sbt-assembly version from 0.7.3 to 0.8.1
  • ISSUE 87: Header Lines support in DelimitedScheme
  • ISSUE 84: Add packTo and unpackTo
  • ISSUE 83: Adds sum and product to Monoid and Ring
  • ISSUE 79: Updates to sbt 0.11.3 to match Travis CI

Version 0.5.3

  • ISSUE 77: Add a scalding multi source tap that has a proper unique identifier

Version 0.5.2

  • ISSUE 74: Upgrade cascading to wip-291 and maple to 0.1.7

Version 0.5.1

  • ISSUE 73: Feature/comparator prop
  • ISSUE 72: Upgrade to cascading wip-288 and maple 0.1.5
  • ISSUE 71: Fixes an issue due to type erasure in KryoHadoopSerialization for scala 2.9.1
  • ISSUE 70: Upgrade scalding to use cascading wip-286.
  • ISSUE 69: Feature/more kryo tests

Version 0.5.0

  • ISSUE 67: Upgrade cascading to wip-281.
  • ISSUE 66: Allow default time zone in DefaultDateRangeJob.
  • ISSUE 65: Fixed the error message thrown by FileSource.validateTaps.
  • ISSUE 62: Kryo Upgrade to 2.04
  • ISSUE 60: Feature/abstract algebra
  • ISSUE 52: Feature/cogroup builder
  • ISSUE 51: Feature/headfix

Version 0.4.1

  • ISSUE 42: Feature/iterable source
  • ISSUE 41: Adds blockJoinWithSmaller to JoinAlgorithms.
  • ISSUE 39: Adding default value to pivot

Version 0.4.0

  • ISSUE 38: Fix bug with hash code collisions of Source objects
  • ISSUE 36: Some cleanups of reduce and Operations
  • ISSUE 35: Split RichPipe join methods into their own trait
  • ISSUE 34: Adds Pivot/Unpivot
  • ISSUE 33: Add pack and unpack methods to RichPipe
  • ISSUE 32: Refactors reducer setting into RichPipe
  • ISSUE 31: Implemented Mode.fileExists
  • ISSUE 28: Simplifies TupleConverter

Version 0.3.5

  • ISSUE 21: move JobTest into main
  • ISSUE 20: Adding a source for the most recent good date path.
Something went wrong with that request. Please try again.