Encode Value class field using the inner type #546

cchantep · 2021-08-08T14:34:04Z

Considering the type below:

class Name(val value: String) extends AnyVal

case class Person(name: Name, age: Int)

Encoding

The name field must be declared as StringType (not StructType(StructField("value", ...))) in the Catalyst representation for Person (see failing test).

Possible fix:

Introduce RecordFieldEncoder typeclass to wrap individual/field TypedEncoder for the implicit resolution required for RecordEncoderFields.
Provide a default RecordFieldEncoder instance for any T: TypedEncoder
Provide a RecordFieldEncoder specific for T <: AnyVal, with a higher resolution priority.

Literal

A literal created using lit function for a Value class instance generate a StructType (see failing test).

Introduce functions.litValue:

frameless.functions.litValue(valueClassInstance) // TypedColumn

cchantep · 2021-08-08T14:34:40Z

dataset/src/test/scala/frameless/RecordEncoderTests.scala

+
+    encoder.jvmRepr shouldBe ObjectType(classOf[Person])
+
+    encoder.catalystRepr shouldBe StructType(Seq(


Test failing on master: https://github.com/typelevel/frameless/pull/546/checks?check_run_id=3423040083#step:8:195

StructType(StructField(name,StructType(StructField(value,StringType,false)),false), StructField(age,IntegerType,false)) was not equal to StructType(StructField(name,StringType,false), StructField(age,IntegerType,false)) (RecordEncoderTests.scala:112)

See possible fix: 0c3f618#diff-e752c554dd3e99366532d2e8ed8a406370715aee66beecb675c4ec38f1f9403aR173

dataset/src/test/scala/frameless/RecordEncoderTests.scala

pomadchin · 2021-08-26T01:06:57Z

Yea, I can see how this can make sense, indeed this can be a nice improvement. I have two questions about the possible fix:

why that should be a wrapped codec? It can live in the TypedEncoder, after the derivation implicit.

  implicit def usingDerivationValueClass[F <: AnyVal, G <: _ :: HNil, H <: _ :: HNil, V]
    (implicit
     i0: LabelledGeneric.Aux[F, G],
     i1: DropUnitValues.Aux[G, H],
     i2: IsHCons.Aux[H, _ <: FieldType[_, V], HNil],
     i3: TypedEncoder[V],
     i4: ClassTag[F]
    ): TypedEncoder[F] = new TypedEncoder[F] {

      def nullable: Boolean = i3.nullable

      def jvmRepr: DataType = i3.catalystRepr

      def catalystRepr: DataType = i3.catalystRepr

      def fromCatalyst(path: Expression): Expression = i3.fromCatalyst(path)

      @inline def toCatalyst(path: Expression): Expression = path
    }

How such codec knows how to construct the value class (the codec above obviously would not be able to construct the value class from string)?

pomadchin · 2021-08-26T02:23:14Z

Oh I see, hm it looks like the scalar case should be handled differently (i got familiar with all issues we have now). It looks like the major issue for now is that we can't establish a relationship between AnyVal and its underlying type.

pomadchin · 2021-08-26T04:17:55Z

That's cool btw, usingDerivationValueClass fixes derivation of calsses with nested value classes. I'm wondering what is exactly wrong with the current derivation (consider the case where AnyVal is a case class)

cchantep · 2021-08-26T08:16:02Z

That's cool btw, usingDerivationValueClass fixes derivation of calsses with nested value classes. I'm wondering what is exactly wrong with the current derivation (consider the case where AnyVal is a case class)

The issue is that at Java reflection level (and so in expressions), a Value class cannot be constructed the same way when used as a field. It's very specific to bytecode optimisation.

The drafted code fixes that : 0c3f618#diff-e752c554dd3e99366532d2e8ed8a406370715aee66beecb675c4ec38f1f9403aR173

Will push thereafter.

pomadchin · 2021-08-26T13:30:40Z

@cchantep I tried to poke around with your code a bit (0c3f618#diff-e752c554dd3e99366532d2e8ed8a406370715aee66beecb675c4ec38f1f9403aR173), and I don't think the following tests pass:

  // passes
  test("Scalar value class") {
    import RecordEncoderTests._

    val encoder = TypedEncoder[Name]

    encoder.jvmRepr shouldBe ObjectType(classOf[Name])

    encoder.catalystRepr shouldBe StructType(Seq(StructField("value", StringType, false)))

    val sqlContext = session.sqlContext
    import sqlContext.implicits._

    TypedDataset
      .createUnsafe[Name](Seq("Foo", "Bar").toDF)(encoder)
      .collect().run() shouldBe Seq(new Name("Foo"), new Name("Bar"))

  }

  // decodes schema but fails on the actual dataset
  //  - Case class with value class field *** FAILED *** (1 second, 117 milliseconds)
  // [info]   java.lang.UnsupportedOperationException: No Encoder found for frameless.RecordEncoderTests.Name
  // [info] - field (class: "frameless.RecordEncoderTests.Name", name: "name")
  // [info] - root class: "frameless.RecordEncoderTests.Person"
  // [info]   at org.apache.spark.sql.catalyst.ScalaReflection$.$anonfun$serializerFor$1(ScalaReflection.scala:591)
  test("Case class with value class field") {
    import RecordEncoderTests._

    val encoder = TypedEncoder[Person]

    encoder.jvmRepr shouldBe ObjectType(classOf[Person])

    encoder.catalystRepr shouldBe StructType(Seq(
      StructField("name", StringType, false),
      StructField("age", IntegerType, false)))

    val sqlContext = session.sqlContext
    import sqlContext.implicits._

    TypedDataset
      .createUnsafe[Person](Seq(Person(new Name("Foo"), 25), Person(new Name("Bar"), 26)).toDF)(encoder)
      .collect().run() shouldBe Seq(Person(new Name("Foo"), 25), Person(new Name("Bar"), 26))
  }

  // decodes schema but fails on the actual dataset
  
  test("Case class with value class as optional field") {
    import RecordEncoderTests._

    // Encode as a User field
    val encoder = TypedEncoder[User]

    encoder.jvmRepr shouldBe ObjectType(classOf[User])

    val expectedPersonStructType = StructType(Seq(
      StructField("id", LongType, false),
      StructField("name", StringType, true)))

    encoder.catalystRepr shouldBe expectedPersonStructType

    val sqlContext = session.sqlContext
    import sqlContext.implicits._

    TypedDataset
      .createUnsafe[User](Seq(User(0, Some(new Name("Foo"))), User(1, None)).toDF)(encoder)
      .collect().run() shouldBe Seq(User(0, Some(new Name("Foo"))), User(1, None))
  }

pomadchin · 2021-08-26T13:32:31Z

btw, @cchantep why is an extra wrapper required? I discovered that putting that implicit directly into the TypedEncoder scope beahves better (though creates an issues with the scalar test)

cchantep · 2021-08-26T13:41:54Z

btw, @cchantep why is an extra wrapper required? I discovered that putting that implicit directly into the TypedEncoder scope beahves better (though creates an issues with the scalar t

Exactly for that reason, handling a Value class as scalar cannot be done in the same way, the having only TypedEncoder cannot distinguish those cases, while the record encoder typeclass can.

pomadchin · 2021-08-26T13:59:47Z

@cchantep got it, than it makes sense to me! I'll look into it more as well to help you (as much as I can) and I think that's really a valuable PR. Even Spark 3.2 can handle AnyVals now! :D

cchantep · 2021-08-29T17:57:24Z

dataset/src/test/scala/frameless/LitTests.scala

+
+    val lorem = new Name("Lorem")
+
+    ds.withColumnReplaced('name, lit(lorem)).


Failing test as lit for Value class generates StructType: failing test)

[info] - support value class *** FAILED *** (89 milliseconds) [info] Array(Q(1,[0,1000000005,6d65726f4c]), Q(2,[0,1000000005,6d65726f4c])) was not equal to List(Q(1,Lorem), Q(2,Lorem)) (LitTests.scala:76)

codecov · 2021-08-29T19:49:47Z

Codecov Report

Merging #546 (ab19018) into master (59caada) will increase coverage by 0.06%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #546      +/-   ##
==========================================
+ Coverage   96.28%   96.35%   +0.06%     
==========================================
  Files          62       63       +1     
  Lines        1051     1071      +20     
  Branches        5        9       +4     
==========================================
+ Hits         1012     1032      +20     
  Misses         39       39

Flag	Coverage Δ
2.12.14	`96.35% <100.00%> (+0.06%)`	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
.../main/scala/frameless/TypedExpressionEncoder.scala	`100.00% <ø> (ø)`
...ataset/src/main/scala/frameless/IsValueClass.scala	`100.00% <100.00%> (ø)`
...taset/src/main/scala/frameless/RecordEncoder.scala	`100.00% <100.00%> (ø)`
dataset/src/main/scala/frameless/TypedColumn.scala	`100.00% <100.00%> (ø)`
...taset/src/main/scala/frameless/functions/Lit.scala	`94.44% <100.00%> (-0.56%)`	⬇️
...t/src/main/scala/frameless/functions/package.scala	`100.00% <100.00%> (ø)`

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 59caada...ab19018. Read the comment docs.

cchantep · 2021-08-31T10:03:12Z

@pomadchin I'm not 100% satisfied, but it can be reviewed

pomadchin · 2021-09-02T12:14:13Z

Hey @cchantep, I will take a look later today or this weekend!

pomadchin

🚀 Left a couple of comments, but I think it looks good and is a solid enhancement.

pomadchin · 2021-09-04T18:20:21Z

dataset/src/main/scala/frameless/IsValueClass.scala

+
+object IsValueClass {
+  /** Provides an evidence `A` is a Value class */
+  implicit def apply[A <: AnyVal, G <: ::[_, HNil], H <: ::[_ <: FieldType[_ <: Symbol, _], HNil]](


IsValueClass is required to disambiguate between Value Classes and Primitives. Being a subtype of AnyVal is not enough i.e. implicitly[Double <:< AnyVal] (that's a comment for myself).

pomadchin · 2021-09-04T18:22:15Z

dataset/src/main/scala/frameless/RecordEncoder.scala

      If(IsNull(path), nullExpr, newExpr)
    }
 }
+
+final class RecordFieldEncoder[T](


I think the RecordFieldEncoder makes sense to me now 👍
I wanted to avoid intoroducing new extra wrappers for derivation, but it is pretty elegant.

dataset/src/main/scala/frameless/functions/Lit.scala

dataset/src/main/scala/frameless/functions/package.scala

pomadchin · 2021-09-04T18:47:24Z

dataset/src/main/scala/frameless/functions/FramelessLit.scala

+import org.apache.spark.sql.types.DataType
+
+@deprecated("Use `Lit[A]`", "0.10.2")
+case class FramelessLit[A](obj: A, encoder: TypedEncoder[A]) extends Expression with NonSQLExpression {


How much sense it makes to keep it in the codebase? It is not used anywhere and I doubt users would be interested in that.

That's a breaking change, so it should only happen for a major version

@cchantep well, I have a feeling that this PR is in general very much binary breaking. It is API compatible though.
Let me set the MIMA plugin in a separate PR!

Removing a public class does break the API.

@cchantep changing public API functions signatures is API breaking as well. I'm wondering what would be the MIMA report now.

That's a question of compromise. Removing this class is not required to fix the current issues, but would break the API (even if we can consider that the user should never directly use this public class), the others (mainly 1) updated signature are strictly required changes.

@cchantep oh yea, but these are public API changes; they break bin compat as well; have you checked the plugin report btw?

I see no reasons why we can't make the major release (and it looks like this PR, with or without this legacy class - is bin incompatible); value classes + Spark 3.2.0 could be good enough for the major version release.

I think @imarios opionion about the versioning scheme is important there as well ¯_(ツ)_/¯ I don't think all frameless minor releases a bin compatible.

dataset/src/test/resources/log4j.properties

pomadchin

🚀 A cool PR; but before merging I'd like to see the MIMA plugin (I can generate it myself) output though, to decide on the FramelessLit fate, and to decide on versioning.

pomadchin · 2021-09-06T11:52:59Z

sbt:frameless> mimaReportBinaryIssues
[info] frameless: mimaPreviousArtifacts is empty, not analyzing binary compatibility.
[info] docs: mimaPreviousArtifacts is empty, not analyzing binary compatibility.
[error] frameless-dataset: Failed binary compatibility check against org.typelevel:frameless-dataset_2.12:0.10.1! Found 6 potential problems
[error]  * static method deriveRecordCons(shapeless.Witness,frameless.TypedEncoder,frameless.RecordEncoderFields)frameless.RecordEncoderFields in interface frameless.RecordEncoderFields's type is different in current version, where it is (shapeless.Witness,frameless.RecordFieldEncoder,frameless.RecordEncoderFields)frameless.RecordEncoderFields instead of (shapeless.Witness,frameless.TypedEncoder,frameless.RecordEncoderFields)frameless.RecordEncoderFields
[error]    filter with: ProblemFilters.exclude[IncompatibleMethTypeProblem]("frameless.RecordEncoderFields.deriveRecordCons")
[error]  * static method deriveRecordLast(shapeless.Witness,frameless.TypedEncoder)frameless.RecordEncoderFields in interface frameless.RecordEncoderFields's type is different in current version, where it is (shapeless.Witness,frameless.RecordFieldEncoder)frameless.RecordEncoderFields instead of (shapeless.Witness,frameless.TypedEncoder)frameless.RecordEncoderFields
[error]    filter with: ProblemFilters.exclude[IncompatibleMethTypeProblem]("frameless.RecordEncoderFields.deriveRecordLast")
[error]  * method deriveRecordLast(shapeless.Witness,frameless.TypedEncoder)frameless.RecordEncoderFields in object frameless.RecordEncoderFields's type is different in current version, where it is (shapeless.Witness,frameless.RecordFieldEncoder)frameless.RecordEncoderFields instead of (shapeless.Witness,frameless.TypedEncoder)frameless.RecordEncoderFields
[error]    filter with: ProblemFilters.exclude[IncompatibleMethTypeProblem]("frameless.RecordEncoderFields.deriveRecordLast")
[error]  * method deriveRecordCons(shapeless.Witness,frameless.TypedEncoder,frameless.RecordEncoderFields)frameless.RecordEncoderFields in object frameless.RecordEncoderFields's type is different in current version, where it is (shapeless.Witness,frameless.RecordFieldEncoder,frameless.RecordEncoderFields)frameless.RecordEncoderFields instead of (shapeless.Witness,frameless.TypedEncoder,frameless.RecordEncoderFields)frameless.RecordEncoderFields
[error]    filter with: ProblemFilters.exclude[IncompatibleMethTypeProblem]("frameless.RecordEncoderFields.deriveRecordCons")
[error]  * static method litAggr(java.lang.Object,frameless.TypedEncoder)frameless.TypedAggregate in class frameless.functions.package does not have a correspondent in current version
[error]    filter with: ProblemFilters.exclude[DirectMissingMethodProblem]("frameless.functions.package.litAggr")
[error]  * method litAggr(java.lang.Object,frameless.TypedEncoder)frameless.TypedAggregate in object frameless.functions.package does not have a correspondent in current version
[error]    filter with: ProblemFilters.exclude[DirectMissingMethodProblem]("frameless.functions.package.litAggr")
[error] stack trace is suppressed; run last dataset / mimaReportBinaryIssues for the full output
[error] (dataset / mimaReportBinaryIssues) Failed binary compatibility check against org.typelevel:frameless-dataset_2.12:0.10.1! Found 6 potential problems
[error] Total time: 2 s, completed Sep 6, 2021 7:51:08 AM

Are these less harmuful than the MissingClass (talking about the potenital removal of the legacy class)? (I'm ok with keeping it, just wondering - mb it makes sense to target to the major release and keep the codebase clean?)

cchantep · 2021-09-06T11:58:00Z

sbt:frameless> mimaReportBinaryIssues
[info] frameless: mimaPreviousArtifacts is empty, not analyzing binary compatibility.
[info] docs: mimaPreviousArtifacts is empty, not analyzing binary compatibility.
[error] frameless-dataset: Failed binary compatibility check against org.typelevel:frameless-dataset_2.12:0.10.1! Found 6 potential problems
[error]  * static method deriveRecordCons(shapeless.Witness,frameless.TypedEncoder,frameless.RecordEncoderFields)frameless.RecordEncoderFields in interface frameless.RecordEncoderFields's type is different in current version, where it is (shapeless.Witness,frameless.RecordFieldEncoder,frameless.RecordEncoderFields)frameless.RecordEncoderFields instead of (shapeless.Witness,frameless.TypedEncoder,frameless.RecordEncoderFields)frameless.RecordEncoderFields
[error]    filter with: ProblemFilters.exclude[IncompatibleMethTypeProblem]("frameless.RecordEncoderFields.deriveRecordCons")
[error]  * static method deriveRecordLast(shapeless.Witness,frameless.TypedEncoder)frameless.RecordEncoderFields in interface frameless.RecordEncoderFields's type is different in current version, where it is (shapeless.Witness,frameless.RecordFieldEncoder)frameless.RecordEncoderFields instead of (shapeless.Witness,frameless.TypedEncoder)frameless.RecordEncoderFields
[error]    filter with: ProblemFilters.exclude[IncompatibleMethTypeProblem]("frameless.RecordEncoderFields.deriveRecordLast")
[error]  * method deriveRecordLast(shapeless.Witness,frameless.TypedEncoder)frameless.RecordEncoderFields in object frameless.RecordEncoderFields's type is different in current version, where it is (shapeless.Witness,frameless.RecordFieldEncoder)frameless.RecordEncoderFields instead of (shapeless.Witness,frameless.TypedEncoder)frameless.RecordEncoderFields
[error]    filter with: ProblemFilters.exclude[IncompatibleMethTypeProblem]("frameless.RecordEncoderFields.deriveRecordLast")
[error]  * method deriveRecordCons(shapeless.Witness,frameless.TypedEncoder,frameless.RecordEncoderFields)frameless.RecordEncoderFields in object frameless.RecordEncoderFields's type is different in current version, where it is (shapeless.Witness,frameless.RecordFieldEncoder,frameless.RecordEncoderFields)frameless.RecordEncoderFields instead of (shapeless.Witness,frameless.TypedEncoder,frameless.RecordEncoderFields)frameless.RecordEncoderFields
[error]    filter with: ProblemFilters.exclude[IncompatibleMethTypeProblem]("frameless.RecordEncoderFields.deriveRecordCons")
[error]  * static method litAggr(java.lang.Object,frameless.TypedEncoder)frameless.TypedAggregate in class frameless.functions.package does not have a correspondent in current version
[error]    filter with: ProblemFilters.exclude[DirectMissingMethodProblem]("frameless.functions.package.litAggr")
[error]  * method litAggr(java.lang.Object,frameless.TypedEncoder)frameless.TypedAggregate in object frameless.functions.package does not have a correspondent in current version
[error]    filter with: ProblemFilters.exclude[DirectMissingMethodProblem]("frameless.functions.package.litAggr")
[error] stack trace is suppressed; run last dataset / mimaReportBinaryIssues for the full output
[error] (dataset / mimaReportBinaryIssues) Failed binary compatibility check against org.typelevel:frameless-dataset_2.12:0.10.1! Found 6 potential problems
[error] Total time: 2 s, completed Sep 6, 2021 7:51:08 AM

Are these less harmuful than the MissingClass (talking about the potenital removal of the legacy class)? (I'm ok with keeping it, just wondering - mb it makes sense to target to the major release and keep the codebase clean?)

MiMa results need to be interpreted.
The real breaking changes (unavoidable) are deriveRecord{Last,Cons}, but they occur at compile-time and should not breaking at runtime.

Anyway if we decide to bump to 0.11, that's fine for me (in which case there is no risk to keep FramelessLit, neither to clean it).

pomadchin · 2021-09-06T12:03:29Z

@cchantep let's do 0.11 than!

cchantep · 2021-09-06T13:21:48Z

@cchantep let's do 0.11 than!

Than https://github.com/typelevel/frameless/pull/449/files could also be included at the same time (no breaking change, but would make a bump more valuable).

``` - support value class *** FAILED *** (101 milliseconds) [info] Array(Q(1,[0,1000000005,6d65726f4c]), Q(2,[0,1000000005,6d65726f4c])) was not equal to List(Q(1,Lorem), Q(2,Lorem)) (LitTests.scala:70) ```

cchantep commented Aug 8, 2021

View reviewed changes

cchantep commented Aug 9, 2021

View reviewed changes

dataset/src/test/scala/frameless/RecordEncoderTests.scala Show resolved Hide resolved

pomadchin added the enhancement label Aug 25, 2021

cchantep linked an issue Aug 26, 2021 that may be closed by this pull request

TypedEncoder not working with UDF (or not properly documented) for class having a field typed with a ValueClass #516

Closed

cchantep marked this pull request as ready for review August 28, 2021 15:32

cchantep force-pushed the feature/valueclass-field branch 4 times, most recently from a4f8950 to 3ff773e Compare August 29, 2021 17:10

cchantep commented Aug 29, 2021

View reviewed changes

cchantep force-pushed the feature/valueclass-field branch from b346330 to 1a83610 Compare August 30, 2021 21:44

pomadchin reviewed Sep 4, 2021

View reviewed changes

pomadchin reviewed Sep 5, 2021

View reviewed changes

dataset/src/test/resources/log4j.properties Outdated Show resolved Hide resolved

cchantep force-pushed the feature/valueclass-field branch 2 times, most recently from 66f29b3 to 7b460a8 Compare September 6, 2021 09:45

pomadchin approved these changes Sep 6, 2021

View reviewed changes

cchantep force-pushed the feature/valueclass-field branch 2 times, most recently from a50a9d1 to ab7610f Compare September 6, 2021 13:56

cchantep added 8 commits September 6, 2021 16:01

Unit test about encoding Value class as dataframe row

53a96bf

Failing test for Value class as field

3bf12ac

Failing test for Value class as optional field

34d631f

Support Value class in record encoder

b40f1a5

Failing test in LitTest:

b74abbf

``` - support value class *** FAILED *** (101 milliseconds) [info] Array(Q(1,[0,1000000005,6d65726f4c]), Q(2,[0,1000000005,6d65726f4c])) was not equal to List(Q(1,Lorem), Q(2,Lorem)) (LitTests.scala:70) ```

Prepare literal refactor, making sure first there is no regression

0ab91ee

Introduce functions.litValue

b43b0d8

IsValueClass evidence

4ae661a

cchantep force-pushed the feature/valueclass-field branch from ab7610f to 9ebeca4 Compare September 6, 2021 14:01

Prepare 0.11

ab19018

cchantep force-pushed the feature/valueclass-field branch from 9ebeca4 to ab19018 Compare September 6, 2021 15:48

cchantep merged commit 3515785 into typelevel:master Sep 6, 2021

cchantep deleted the feature/valueclass-field branch September 6, 2021 16:36

cchantep mentioned this pull request Sep 7, 2021

AnyVal blows up TypedEncoder / Injection #161

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Encode Value class field using the inner type #546

Encode Value class field using the inner type #546

cchantep commented Aug 8, 2021 •

edited

Loading

cchantep Aug 8, 2021 •

edited

Loading

cchantep Aug 9, 2021

pomadchin commented Aug 26, 2021 •

edited

Loading

pomadchin commented Aug 26, 2021 •

edited

Loading

pomadchin commented Aug 26, 2021

cchantep commented Aug 26, 2021

pomadchin commented Aug 26, 2021 •

edited

Loading

pomadchin commented Aug 26, 2021

cchantep commented Aug 26, 2021 •

edited

Loading

pomadchin commented Aug 26, 2021

cchantep Aug 29, 2021

codecov bot commented Aug 29, 2021 •

edited

Loading

cchantep commented Aug 31, 2021

pomadchin commented Sep 2, 2021

pomadchin left a comment

pomadchin Sep 4, 2021

pomadchin Sep 4, 2021

pomadchin Sep 4, 2021

cchantep Sep 4, 2021

pomadchin Sep 4, 2021

cchantep Sep 5, 2021

pomadchin Sep 5, 2021

cchantep Sep 6, 2021

pomadchin Sep 6, 2021

pomadchin Sep 6, 2021

pomadchin left a comment •

edited

Loading

pomadchin commented Sep 6, 2021 •

edited

Loading

cchantep commented Sep 6, 2021

pomadchin commented Sep 6, 2021

cchantep commented Sep 6, 2021


		encoder.jvmRepr shouldBe ObjectType(classOf[Person])

		encoder.catalystRepr shouldBe StructType(Seq(


		val lorem = new Name("Lorem")

		ds.withColumnReplaced('name, lit(lorem)).

Encode Value class field using the inner type #546

Encode Value class field using the inner type #546

Conversation

cchantep commented Aug 8, 2021 • edited Loading

Encoding

Literal

cchantep Aug 8, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

pomadchin commented Aug 26, 2021 • edited Loading

pomadchin commented Aug 26, 2021 • edited Loading

pomadchin commented Aug 26, 2021

cchantep commented Aug 26, 2021

pomadchin commented Aug 26, 2021 • edited Loading

pomadchin commented Aug 26, 2021

cchantep commented Aug 26, 2021 • edited Loading

pomadchin commented Aug 26, 2021

Choose a reason for hiding this comment

codecov bot commented Aug 29, 2021 • edited Loading

Codecov Report

cchantep commented Aug 31, 2021

pomadchin commented Sep 2, 2021

pomadchin left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

pomadchin left a comment • edited Loading

Choose a reason for hiding this comment

pomadchin commented Sep 6, 2021 • edited Loading

cchantep commented Sep 6, 2021

pomadchin commented Sep 6, 2021

cchantep commented Sep 6, 2021

cchantep commented Aug 8, 2021 •

edited

Loading

cchantep Aug 8, 2021 •

edited

Loading

pomadchin commented Aug 26, 2021 •

edited

Loading

pomadchin commented Aug 26, 2021 •

edited

Loading

pomadchin commented Aug 26, 2021 •

edited

Loading

cchantep commented Aug 26, 2021 •

edited

Loading

codecov bot commented Aug 29, 2021 •

edited

Loading

pomadchin left a comment •

edited

Loading

pomadchin commented Sep 6, 2021 •

edited

Loading