spark-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From cloud-fan <...@git.apache.org>
Subject [GitHub] spark pull request #16386: [SPARK-18352][SQL] Support parsing multiline json...
Date Fri, 17 Feb 2017 04:50:42 GMT
Github user cloud-fan commented on a diff in the pull request:

    https://github.com/apache/spark/pull/16386#discussion_r101684365
  
    --- Diff: sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/json/JsonSuite.scala
---
    @@ -1802,4 +1806,142 @@ class JsonSuite extends QueryTest with SharedSQLContext with TestJsonData
{
         val df2 = spark.read.option("PREfersdecimaL", "true").json(records)
         assert(df2.schema == schema)
       }
    +
    +  test("SPARK-18352: Parse normal multi-line JSON files (compressed)") {
    +    withTempPath { dir =>
    +      val path = dir.getCanonicalPath
    +      primitiveFieldAndType
    +        .toDF("value")
    +        .write
    +        .option("compression", "GzIp")
    +        .text(path)
    +
    +      assert(new File(path).listFiles().exists(_.getName.endsWith(".gz")))
    +
    +      val jsonDF = spark.read.option("wholeFile", true).json(path)
    +      val jsonDir = new File(dir, "json").getCanonicalPath
    +      jsonDF.coalesce(1).write
    +        .option("compression", "gZiP")
    +        .json(jsonDir)
    +
    +      assert(new File(jsonDir).listFiles().exists(_.getName.endsWith(".json.gz")))
    +
    +      val originalData = spark.read.json(primitiveFieldAndType)
    +      checkAnswer(jsonDF, originalData)
    +      checkAnswer(spark.read.schema(originalData.schema).json(jsonDir), originalData)
    +    }
    +  }
    +
    +  test("SPARK-18352: Parse normal multi-line JSON files (uncompressed)") {
    +    withTempPath { dir =>
    +      val path = dir.getCanonicalPath
    +      primitiveFieldAndType
    +        .toDF("value")
    +        .write
    +        .text(path)
    +
    +      val jsonDF = spark.read.option("wholeFile", true).json(path)
    +      val jsonDir = new File(dir, "json").getCanonicalPath
    +      jsonDF.coalesce(1).write.json(jsonDir)
    +
    +      val compressedFiles = new File(jsonDir).listFiles()
    +      assert(compressedFiles.exists(_.getName.endsWith(".json")))
    +
    +      val originalData = spark.read.json(primitiveFieldAndType)
    +      checkAnswer(jsonDF, originalData)
    +      checkAnswer(spark.read.schema(originalData.schema).json(jsonDir), originalData)
    +    }
    +  }
    +
    +  test("SPARK-18352: Expect one JSON document per file") {
    +    // the json parser terminates as soon as it sees a matching END_OBJECT or END_ARRAY
token.
    +    // this might not be the optimal behavior but this test verifies that only the first
value
    +    // is parsed and the rest are discarded.
    +
    +    // alternatively the parser could continue parsing following objects, which may further
reduce
    +    // allocations by skipping the line reader entirely
    +
    +    withTempPath { dir =>
    +      val path = dir.getCanonicalPath
    +      spark
    +        .createDataFrame(Seq(Tuple1("{}{invalid}")))
    +        .coalesce(1)
    +        .write
    +        .text(path)
    +
    +      val jsonDF = spark.read.option("wholeFile", true).json(path)
    +      // no corrupt record column should be created
    +      assert(jsonDF.schema === StructType(Seq()))
    +      // only the first object should be read
    +      assert(jsonDF.count() === 1)
    +    }
    +  }
    +
    +  test("SPARK-18352: Handle multi-line corrupt documents (PERMISSIVE)") {
    +    withTempPath { dir =>
    +      val path = dir.getCanonicalPath
    +      val corruptRecordCount = additionalCorruptRecords.count().toInt
    +      assert(corruptRecordCount === 5)
    +
    +      additionalCorruptRecords
    --- End diff --
    
    I have a more robust solution:
    ```
    additionalCorruptRecords.collect().zipWithIndex.foreach {
      case (str, index) =>
        Seq(str).toDF("value").write.text(new File(dir, index + ".json").getCanonicalPath)
    }
    ... // dir will have 5 files: 0.json, 1.json, ... 4.json
    ```


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org


Mime
View raw message