Elasticsearch "mapper_parsing_exception" in CI

Elasticsearch could not parse a document field into the type declared in the index mapping. A string sent where a number or date is mapped, or a bad date format, produces mapper_parsing_exception on the index request. In CI this often surfaces because a fresh cluster inferred a mapping from the first document.

What this error means

An index or bulk request fails with "mapper_parsing_exception" and "failed to parse field [X] of type [Y]", naming the field and the expected type.

elasticsearch

{"error":{"type":"mapper_parsing_exception","reason":"failed to parse field [age]
of type [long] in document with id '1'. ...","caused_by":{"type":"illegal_argument_exception",
"reason":"For input string: \"twenty\""}},"status":400}

Common causes

A field value does not match the mapped type

The mapping declares (or inferred) a numeric or date type, but the document sends a value that cannot be parsed into it.

Dynamic mapping guessed the wrong type

On a fresh CI cluster the first indexed document sets the dynamic mapping; a later document with a different shape then fails to parse.

How to fix it

Define an explicit mapping before indexing

Create the index with an explicit mapping instead of relying on dynamic inference.
Match the field types to the data your tests actually index.
Run this as a setup step so every run starts from the same mapping.

Terminal

curl -X PUT "http://localhost:9200/products" -H 'Content-Type: application/json' -d '
{"mappings":{"properties":{"age":{"type":"integer"},"created":{"type":"date"}}}}'

Send values that match the mapping

Fix the document so each field parses into its mapped type, for example a numeric age and an ISO-8601 date.

How to prevent it

Create indices with explicit mappings in a CI setup step.
Do not rely on dynamic mapping order, which depends on which document lands first.
Validate fixture data against the mapping before indexing.