AWS solutions documentation change
Summary
Enhanced documentation of Solr to OpenSearch schema translation. Expanded field type mapping table, added details about date formats, dynamic field handling, copyField resolution, metadata migration settings, and fallback behavior for unrecognized types.
Security assessment
The changes are purely functional improvements to schema translation documentation. No security vulnerabilities, exploits, or security features are mentioned. Changes focus on mapping accuracy, data type handling, and migration workflow enhancements without security implications.
Diff
diff --git a/solutions/latest/migration-assistant-for-amazon-opensearch-service/solr-schema-translation.md b/solutions/latest/migration-assistant-for-amazon-opensearch-service/solr-schema-translation.md index b6c8d45ca..90cfd3999 100644 --- a//solutions/latest/migration-assistant-for-amazon-opensearch-service/solr-schema-translation.md +++ b//solutions/latest/migration-assistant-for-amazon-opensearch-service/solr-schema-translation.md @@ -9 +9 @@ -The SolrReader maps common Solr field types to their OpenSearch equivalents automatically. The following table lists the built-in translations. +The SolrReader parses the collection schema from the backup (`managed-schema.xml`, `managed-schema`, or `schema.xml`) and converts explicit fields, dynamic field patterns, and `copyField` destinations into OpenSearch mappings. The following table lists the built-in type translations. The converter first checks the Solr field type name, then falls back to the field type’s Java class when the name is custom. @@ -11 +11 @@ The SolrReader maps common Solr field types to their OpenSearch equivalents auto -Solr field type | OpenSearch mapping +Solr field type name or class | OpenSearch mapping @@ -13,10 +13,11 @@ Solr field type | OpenSearch mapping -`solr.TextField` | `text` -`solr.StrField` | `keyword` -`solr.IntPointField` | `integer` -`solr.LongPointField` | `long` -`solr.FloatPointField` | `float` -`solr.DoublePointField` | `double` -`solr.BoolField` | `boolean` -`solr.DatePointField` | `date` - -Field types that are not in this table — for example, custom or third-party Solr field types — are not translated automatically. Review your `schema.xml` before you migrate, and validate the resulting mappings on the target after backfill completes (see [Configure and run the backfill workflow](./solr-run-workflow.html)). For field types outside this set, plan to define equivalent OpenSearch mappings as part of your migration. +`string`, `strings`, `solr.StrField`, `solr.UUIDField` | `keyword` +`text`, `text_general`, `text_en`, `text_ws`, `solr.TextField` | `text` +`pint`, `pints`, `int`, `tint`, `solr.IntPointField`, `solr.TrieIntField` | `integer` +`plong`, `plongs`, `long`, `tlong`, `solr.LongPointField`, `solr.TrieLongField` | `long` +`pfloat`, `pfloats`, `float`, `tfloat`, `solr.FloatPointField`, `solr.TrieFloatField` | `float` +`pdouble`, `pdoubles`, `double`, `tdouble`, `solr.DoublePointField`, `solr.TrieDoubleField` | `double` +`boolean`, `booleans`, `solr.BoolField` | `boolean` +`pdate`, `pdates`, `date`, `tdate`, `solr.DatePointField`, `solr.TrieDateField` | `date` +`binary`, `solr.BinaryField` | `binary` + +Date mappings are created with `strict_date_optional_time||epoch_millis` so migrated values can be indexed from either ISO-8601 strings or epoch-millis values. Dynamic fields such as ` **_s_ ** become OpenSearch dynamic templates, with type-aware matching so dotted Solr field names do not accidentally type intermediate object paths. `copyField` destinations are added to the mapping when the destination is not already an explicit field; the converter resolves their type from matching dynamic field patterns when possible and otherwise maps them as `text`. Internal fields whose names start with `_` are skipped, except for the Solr `id` field.` Metadata migration also creates basic index settings for each Solr collection or core. `index.number_of_shards` is set from the shard count discovered in the backup, and `index.number_of_replicas` defaults to `1`. Review these settings at the metadata evaluation gate and use a metadata transform if your target needs a different replica count or other index settings. Field types that are not recognized by name or Java class fall back to `text`. Review your Solr schema before you migrate, and validate the resulting mappings on the target after backfill completes (see [Configure and run the backfill workflow](./solr-run-workflow.html)). For custom or third-party field types, plan to define or adjust equivalent OpenSearch mappings as part of your migration. @@ -30 +31 @@ To use the Amazon Web Services Documentation, Javascript must be enabled. Please -How Solr migration differs +Document reconstruction