voonhous commented on code in PR #18837:
URL: https://github.com/apache/hudi/pull/18837#discussion_r3340407323
##########
hudi-trino-plugin/src/main/java/io/trino/plugin/hudi/HudiUtil.java:
##########
@@ -397,4 +413,62 @@ public static Schema
getLatestTableSchema(HoodieTableMetaClient metaClient, Stri
throw new TrinoException(HUDI_FILESYSTEM_ERROR, e);
}
}
+
+ public static List<HiveColumnHandle> getOrderingColumnHandles(Table table,
TypeManager typeManager, Lazy<HoodieTableMetaClient> lazyMetaClient,
HiveTimestampPrecision timestampPrecision)
+ {
+ RecordMergeMode recordMergeMode =
lazyMetaClient.get().getTableConfig().getRecordMergeMode();
+ if (Objects.isNull(recordMergeMode) ||
recordMergeMode.equals(RecordMergeMode.COMMIT_TIME_ORDERING)) {
+ // if commit time ordering is enabled, return empty list
+ return Collections.emptyList();
+ }
+
+ ImmutableList.Builder<HiveColumnHandle> columns =
ImmutableList.builder();
+ List<String> orderingColumnNames =
lazyMetaClient.get().getTableConfig().getOrderingFields();
+
+ int hiveColumnIndex = 0;
+ for (Column field : table.getDataColumns()) {
+ // ignore unsupported types rather than failing
+ if (orderingColumnNames.contains(field.getName())) {
+ HiveType hiveType = field.getType();
+ if (typeSupported(hiveType.getTypeInfo(),
table.getStorage().getStorageFormat())) {
+ columns.add(createBaseColumn(field.getName(),
hiveColumnIndex, hiveType, getType(hiveType, typeManager, timestampPrecision),
REGULAR, field.getComment()));
+ }
+ }
+ hiveColumnIndex++;
+ }
+
+ return columns.build();
+ }
+
+ /**
+ * Converts the given {@link HoodiePairData} into a {@link Map}.
+ * <p>
+ * Special handling is applied for null keys:
+ * <ul>
+ * <li>If a key is null, it is stored in the map as a {@code null}
entry.</li>
+ * <li>If multiple entries share the same key (including null), the
latest value overwrites the previous one.</li>
+ * </ul>
+ *
+ * @param pairData the HoodiePairData containing key-value pairs
+ * @param <K> the type of keys maintained by the resulting map
+ * @param <V> the type of mapped values
+ * @return a {@link Map} containing all key-value pairs from the input data
+ */
+ public static <K, V> Map<K, V> collectAsMap(HoodiePairData<K, V> pairData)
+ {
+ // Map each pair to (Option<Pair.key>, V) to handle null keys uniformly
+ // If there are multiple entries sharing the same key, use the
incoming one
+ return pairData.mapToPair(pair ->
+ Pair.of(
Review Comment:
Addressed in ebc5077. Dropped the `mapToPair` +
`Option.ofNullable`/`orElse(null)` round-trip and collect the pairs directly
(HashMap accepts null keys; later entries overwrite earlier ones). Also removed
the now-unused `Option` and `Pair` imports.
--
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]
