sandip-db commented on code in PR #43319:
URL: https://github.com/apache/spark/pull/43319#discussion_r1354166034
##########
sql/core/src/test/resources/test-data/xml-resources/root-level-value.xml:
##########
@@ -0,0 +1,9 @@
+<?xml version="1.0"?>
+<ROWSET>
+ <ROW>value1</ROW>
+ <ROW attr="attr1">value2</ROW>
+ <ROW>
+ value3
+ <!-- this is a comment-->
+ </ROW>
+</ROWSET>
Review Comment:
pls add and test with the following ROW:
```
<ROW>4<tag>5</tag></ROW>
<ROW><tag>6</tag>7</ROW>
<ROW attr="8"></ROW>
```
4 and 7 should be skipped.
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/xml/XmlInferSchema.scala:
##########
@@ -219,11 +219,27 @@ private[sql] object XmlInferSchema {
dataTypes += inferredType
nameToDataType += (field -> dataTypes)
+ case c: Characters if !c.isWhiteSpace =>
+ // This can be an attribute-only object
+ val valueTagType = inferFrom(c.getData, options)
+ nameToDataType += options.valueTag -> ArrayBuffer(valueTagType)
+
case _: EndElement =>
shouldStop = StaxXmlParserUtils.checkEndElement(parser)
case _ => // do nothing
}
+ }
+ // A structure object is an attribute-only element
+ // if it only consists of attributes and valueTags.
+ // If not, we will remove the valueTag field from the schema
+ val attributesOnly = nameToDataType.forall {
+ case (fieldName, dataTypes) =>
+ dataTypes.length == 1 &&
+ (fieldName == options.valueTag ||
fieldName.startsWith(options.attributePrefix))
Review Comment:
This is fragile as an element name can begin with the
`options.attributePrefix`.
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/xml/XmlInferSchema.scala:
##########
@@ -219,11 +219,27 @@ private[sql] object XmlInferSchema {
dataTypes += inferredType
nameToDataType += (field -> dataTypes)
+ case c: Characters if !c.isWhiteSpace =>
+ // This can be an attribute-only object
+ val valueTagType = inferFrom(c.getData, options)
+ nameToDataType += options.valueTag -> ArrayBuffer(valueTagType)
Review Comment:
Can this be added outside the `while` loop after examining that no child
elements were present?
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/xml/StaxXmlParser.scala:
##########
@@ -122,7 +122,12 @@ class StaxXmlParser(
}
val parser = StaxXmlParserUtils.filteredReader(xml)
val rootAttributes = StaxXmlParserUtils.gatherRootAttributes(parser)
- Some(convertObject(parser, schema, options, rootAttributes))
+ // A structure object is an attribute-only element
+ // if it only consists of attributes and valueTags.
+ val isRootAttributesOnly = schema.fields.forall { f =>
+ f.name == options.valueTag ||
f.name.startsWith(options.attributePrefix)
+ }
+ Some(convertObject(parser, schema, options, rootAttributes,
isRootAttributesOnly))
Review Comment:
Couldn't we just pass `true` instead of `isRootAttributesOnly` here? There
is a check for `nameToIndex.get(options.valueTag)` below anyway.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
