STRICT_DUPLICATE_DETECTION and FAIL_ON_READING_DUP_TREE_KEY allow for duplications, when they come from child-nodes

[bjuergens]

Search before asking

• I searched in the issues and found nothing similar.

(#237 describes FAIL_ON_READING_DUP_TREE_KEY )

Describe the bug

When mapping to a non-list, like a string, I expect to get an exception, when there is a duplicate value. As it is, the first value just disappears, with no possibility to even know there was something wrong to begin with. As it stands, "FAIL_ON_READING_DUP_TREE_KEY" only picks up on duplicate Attributes, which is already broken XML to begin with.

example:

<myXml>
    <value>aaa</value>
    <value>bbb</value>
</myXml> ```

with this class 
```kotlin
data class MyXml(val value: String)

here MyXml.value receives a duplicate value, which FAIL_ON_READING_DUP_TREE_KEY will not detect.

I think the solution would either to add this to the "known limitation" section in the readme, or to fix the bug.

Is there a workaround for this bug, so that I can find out, when multiple values gets mapped to the same property? (And no, I don't want to turn every single property into a list, and then manually check if size<=1. But maybe there is something like a global flag, that adds a custom setter to all properties, that checks if there was already a value? Or maybe there is some way to add a custom Feature to add this check?)

Version Information

2.18.2

Reproduction

import com.fasterxml.jackson.core.JsonParser
import com.fasterxml.jackson.databind.DeserializationFeature
import com.fasterxml.jackson.dataformat.xml.XmlMapper
import com.fasterxml.jackson.module.kotlin.kotlinModule
import org.testng.annotations.Test
import kotlin.test.assertEquals
data class MyXml(val value: String)

class DuplicateTest {

    @Test(expectedExceptions = [com.fasterxml.jackson.core.JsonParseException::class])
    fun duplicateNodeTestWorksWithAttributes() {
        val mapper = XmlMapper.builder().addModule(kotlinModule()).enable(DeserializationFeature.FAIL_ON_READING_DUP_TREE_KEY).enable(JsonParser.Feature.STRICT_DUPLICATE_DETECTION).build()
        mapper.readValue("""<myXml value="aaa" value="bbb" />""", MyXml::class.java)
    }
    
    @Test(expectedExceptions = [com.fasterxml.jackson.core.JsonParseException::class]) 
    fun duplicateNodeTestDoesntWorkWithChildren() {
        val mapper = XmlMapper.builder().addModule(kotlinModule()).enable(DeserializationFeature.FAIL_ON_READING_DUP_TREE_KEY).enable(JsonParser.Feature.STRICT_DUPLICATE_DETECTION).build()
        val result = mapper.readValue("""<myXml>
                    <value>aaa</value>
                    <value>bbb</value>
                </myXml> """, MyXml::class.java)

        assertEquals("bbb", result.value)
        assertEquals("aaa", result.value)
    }
}

Expected behavior

I expect "STRICT_DUPLICATE_DETECTION" to give me an exception, whenever an existing value gets overwritten, and if the target-property is no list.

Additional context

I am no jackson-XML professional, so odds are, that I overlooked something obvious. I that is the case, I am very sorry.

(edit: formatting)

[JooHyukKim]

@bjuergens I don't understand how this part works. Is it possible?
Or these assertions are only placeholder?

        assertEquals("bbb", result.value)
        assertEquals("aaa", result.value)

Also, you may want to read through DeserializationFeature.FAIL_ON_READING_DUP_TREE_KEY documentation again, to verify it applies to your case.

[cowtowncoder]

Like @JooHyukKim suggests, DeserializationFeature.FAIL_ON_READING_DUP_TREE_KEY ONLY applies when reading content as JsonNode:

     * Feature that determines what happens when reading JSON content into tree
     * ({@link JsonNode} and a duplicate key
     * is encountered (property name that was already seen for the JSON Object).
     * If enabled, {@link JsonMappingException} will be thrown; if disabled, no exception
     * is thrown and the new (later) value overwrites the earlier value.
     *<p>
     * Note that this property does NOT affect other aspects of data-binding; that is,
     * no detection is done with respect to POJO properties or {@link java.util.Map}
     * keys. New features may be added to control additional cases.

so test is not valid reproduction.

It is possible this might not work, but test would need to try reading XML content as JsonNode (XmlMapper.readTree(...) f.ex).

[bjuergens]

@bjuergens I don't understand how this part works. Is it possible? Or these assertions are only placeholder?

Is it possible?

no.

Or these assertions are only placeholder?

yes.

[bjuergens]

Like @JooHyukKim suggests, DeserializationFeature.FAIL_ON_READING_DUP_TREE_KEY ONLY applies when reading content as JsonNode:

Is there some other way to prevent silent data loss while using FasterXML?

[JooHyukKim]

Since your POJO will have setter somewhere.
If field value already set, throw error?

[bjuergens]

If field value already set, throw error?

Yes, this should work for individual fields.

Would it be able to activate such behavior across all models at once? something like what throwOnRepeatedElement does for pdvrieze/xmlutil?

[cowtowncoder]

If field value already set, throw error?

Yes, this should work for individual fields.

Would it be able to activate such behavior across all models at once? something like what throwOnRepeatedElement does for pdvrieze/xmlutil?

Unfortunately, no: Jackson does not keep track of bound objects (results) or which properties have been set.
FAIL_ON_READING_DUP_TREE_KEY works since deserializer for JsonNode does have this state, so it can be implemented.

Although what might work is using DeserializationFeature.FAIL_ON_READING_DUP_TREE_KEY with XmlMapper.readTree() and then using mapper.treeToValue(node, TargetType.class)

Another note on something I missed originally: JsonParser.Feature.STRICT_DUPLICATE_DETECTION is unfortunately not yet implemented by XML backend. If it was, that would have caught duplicate.
I'll see if there is an issue for missing support; if not, will create one. (that doesn't mean it gets implemented immediately but raises visibility if someone wanted to tackle it).

[cowtowncoder]

Ok: issue -- #114 -- does exist, all the way since 2014. :-/

[JooHyukKim]

@cowtowncoder is there a similar usage I can follow, to implement JsonParser.Feature.STRICT_DUPLICATE_DETECTION on module extension?

[cowtowncoder]

@JooHyukKim CBOR and Smile backends also support it (there was bug reported FasterXML/jackson-dataformats-binary#312 for 2.14), so maybe having a look at CBORParser.java and SmileParserBase.java would help?
It also looks like YAMLParser supports duplicate detection since there is yaml/src/test/java/com/fasterxml/jackson/dataformat/yaml/deser/ParserDupHandlingTest.java.

I was also thinking that one could use a tool like Claude Code to possibly do this: or at least create test cases.
GenAI tools are pretty good at learning and using patterns in existing code -- not sure how easy it'd be to point to CBOR or Smile sources, to use for XML. Just an idea.