[DRILL-7060] Support JsonParser Feature 'ALLOW_BACKSLASH_ESCAPING_ANY_CHARACTER' in JsonReader

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.15.0, 1.16.0
• Fix Version/s: 1.16.0
• Component/s: Storage - JSON
• Labels: ready-to-commit
• Assignee: Abhishek Girish
• Reporter: Abhishek Girish
• Reviewer: Sorabh Hamirwasia
• Created: 27/Feb/19 18:26
• Updated: 27/Mar/20 00:08
• Resolved: 04/Mar/19 08:20

Description

Some JSON files may have strings with backslashes - which are read as escape characters. By default only standard escape characters are allowed. So querying such files would fail. For example see:

Data

{"file":"C:\Sfiles\escape.json"}

Error

(com.fasterxml.jackson.core.JsonParseException) Unrecognized character escape 'S' (code 83)
 at [Source: (org.apache.drill.exec.store.dfs.DrillFSDataInputStream); line: 1, column: 178]
    com.fasterxml.jackson.core.JsonParser._constructError():1804
    com.fasterxml.jackson.core.base.ParserMinimalBase._reportError():663
    com.fasterxml.jackson.core.base.ParserMinimalBase._handleUnrecognizedCharacterEscape():640
    com.fasterxml.jackson.core.json.UTF8StreamJsonParser._decodeEscaped():3243
    com.fasterxml.jackson.core.json.UTF8StreamJsonParser._skipString():2537
    com.fasterxml.jackson.core.json.UTF8StreamJsonParser.nextToken():683
    org.apache.drill.exec.vector.complex.fn.JsonReader.writeData():342
    org.apache.drill.exec.vector.complex.fn.JsonReader.writeDataSwitch():298
    org.apache.drill.exec.vector.complex.fn.JsonReader.writeToVector():246
    org.apache.drill.exec.vector.complex.fn.JsonReader.write():205
    org.apache.drill.exec.store.easy.json.JSONRecordReader.next():216
    org.apache.drill.exec.physical.impl.ScanBatch.internalNext():223
...
...

Issue Links

• links to: GitHub Pull Request #1663