Add file abstraction types, core utilities, and inline I/O for Hudi connector by voonhous · Pull Request #28518 · trinodb/trino

voonhous · 2026-03-04T03:28:31Z

Description

This PR adds foundational building blocks to the Hudi connector needed for full COW and MOR table read support.

File abstraction layer (io.trino.plugin.hudi.file)

Introduces HudiFile as a common interface for Hudi files, with two concrete implementations:

HudiBaseFile - a Trino-native wrapper around HoodieBaseFile, holding path, file name, size, modification time, and a byte-range [start, length] for split-aligned reads.
HudiLogFile - same structure wrapping HoodieLogFile, used for MOR delta log files.

Both are Jackson-serialisable so they can be round-tripped through Trino's split serialisation mechanism.

Utilities (io.trino.plugin.hudi.util)

HudiTableTypeUtils - maps Hive input format class names (both legacy com.uber.hoodie and current org.apache.hudi namespaces) to HoodieTableType (COPY_ON_WRITE / MERGE_ON_READ), replacing ad-hoc string comparisons scattered across the plugin.
HudiAvroSerializer - bidirectional conversion between Avro GenericRecord values and Trino Block/Page values, covering all standard SQL types (boolean, integer family, real/double, decimal, date, timestamp, timestamp with time zone, varchar, char, varbinary, array, map, row). Required for deserialising Hudi log-file blocks during MOR compaction reads.
TupleDomainUtils - translates Trino TupleDomain predicates into column-level value lists for use against Hudi index metadata (Bloom filter, column stats, record-level index). Handles range decomposition, null handling, and statistics overflow.

Inline filesystem I/O (io.trino.plugin.hudi.io / storage)

Hudi log files can embed smaller files using the inlinefs:// URI scheme (InLineFSUtils). Added:

InlineSeekableDataInputStream - wraps a TrinoInputStream with offset/length bounds so that SeekableDataInputStream contract is satisfied for inline payloads.
TrinoHudiInlineStorage - a HoodieStorage implementation that delegates inline-scheme paths to a bounded sub-read of the outer TrinoHudiStorage, and routes all other operations back to the base storage. Required for reading Hudi log files that contain embedded column-stats or Bloom-filter metadata blocks.

HudiUtil & minor cleanup

Extends HudiUtil with additional helpers, and adjusts HudiSplitSource and TimelineTable to use the new abstractions.

NOTE: This PR is part [1/4] of efforts to upstream Hudi optimizations to opensource.

Additional context and related issues

The Hudi connector currently supports only Copy-on-Write tables and relies on Hadoop FileSystem for I/O. This work is part of a broader effort to:

Remove the Hadoop FileSystem dependency in favour of Trino's own TrinoFileSystem abstraction.
Enable Merge-on-Read table reads by implementing the full log-file compaction pipeline.

The inline filesystem support (InLineFSUtils, inlinefs:// scheme) is a Hudi-internal mechanism used when storing small metadata (Bloom filters, column statistics) as byte ranges inside a log file rather than as separate files. Without TrinoHudiInlineStorage, the Hudi library will fall back to a Hadoop FileSystem call and fail when Hadoop is not present.

TupleDomainUtils mirrors logic from TupleDomainParquetPredicate but operates on string-keyed column-stats maps returned by the Hudi metadata table, rather than Parquet row-group statistics.

Release notes

(x) This is not user-visible or is docs only, and no release notes are required.

cla-bot · 2026-03-04T03:28:35Z

Thank you for your pull request and welcome to the Trino community. We require contributors to sign our Contributor License Agreement, and we don't seem to have you on file. Continue to work with us on the review and improvements in this PR, and submit the signed CLA to cla@trino.io. Photos, scans, or digitally-signed PDF files are all suitable. Processing may take a few days. The CLA needs to be on file before we merge your changes. For more information, see https://github.com/trinodb/cla

cla-bot · 2026-03-04T03:53:00Z

Thank you for your pull request and welcome to the Trino community. We require contributors to sign our Contributor License Agreement, and we don't seem to have you on file. Continue to work with us on the review and improvements in this PR, and submit the signed CLA to cla@trino.io. Photos, scans, or digitally-signed PDF files are all suitable. Processing may take a few days. The CLA needs to be on file before we merge your changes. For more information, see https://github.com/trinodb/cla

cla-bot · 2026-03-04T04:03:38Z

Thank you for your pull request and welcome to the Trino community. We require contributors to sign our Contributor License Agreement, and we don't seem to have you on file. Continue to work with us on the review and improvements in this PR, and submit the signed CLA to cla@trino.io. Photos, scans, or digitally-signed PDF files are all suitable. Processing may take a few days. The CLA needs to be on file before we merge your changes. For more information, see https://github.com/trinodb/cla

cla-bot · 2026-03-04T04:08:41Z

Thank you for your pull request and welcome to the Trino community. We require contributors to sign our Contributor License Agreement, and we don't seem to have you on file. Continue to work with us on the review and improvements in this PR, and submit the signed CLA to cla@trino.io. Photos, scans, or digitally-signed PDF files are all suitable. Processing may take a few days. The CLA needs to be on file before we merge your changes. For more information, see https://github.com/trinodb/cla

cla-bot · 2026-03-04T04:44:11Z

Thank you for your pull request and welcome to the Trino community. We require contributors to sign our Contributor License Agreement, and we don't seem to have you on file. Continue to work with us on the review and improvements in this PR, and submit the signed CLA to cla@trino.io. Photos, scans, or digitally-signed PDF files are all suitable. Processing may take a few days. The CLA needs to be on file before we merge your changes. For more information, see https://github.com/trinodb/cla

Copilot

Pull request overview

This PR adds core building blocks for expanding the Trino Hudi connector toward full Copy-on-Write (COW) and Merge-on-Read (MOR) read support, including file abstractions, predicate/statistics utilities, and inlinefs (inlinefs://) I/O plumbing needed by Hudi log files.

Changes:

Introduce Hudi file abstractions (HudiFile, HudiBaseFile, HudiLogFile) suitable for split serialization.
Add utility components (TupleDomainUtils, HudiTableTypeUtils, HudiAvroSerializer) to support predicate handling, table type detection, and Avro<->Trino conversions.
Add inlinefs storage and bounded seekable input stream support (TrinoHudiInlineStorage, InlineSeekableDataInputStream) and wire inlinefs dispatch in TrinoHudiStorage.

Reviewed changes

Copilot reviewed 20 out of 20 changed files in this pull request and generated 9 comments.

Show a summary per file

File	Description
plugin/trino-hudi/src/test/java/io/trino/plugin/hudi/util/TestTupleDomainUtilsTest.java	Adds unit tests for TupleDomainUtils column/reference helpers and IN/EQUAL-only detection.
plugin/trino-hudi/src/test/java/io/trino/plugin/hudi/util/TestTupleDomainUtilsExtendedNullFilterTest.java	Adds unit tests for simple null-check detection in TupleDomainUtils.
plugin/trino-hudi/src/test/java/io/trino/plugin/hudi/storage/TestTrinoHudiStorage.java	Updates expected `getUri()` behavior for TrinoHudiStorage.
plugin/trino-hudi/src/test/java/io/trino/plugin/hudi/io/TestInlineSeekableDataInputStream.java	Adds unit tests for inline seekable stream initialization/seek/read behavior.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/util/TupleDomainUtils.java	New tuple-domain helpers for referenced columns, null checks, stats evaluation, and record-key construction.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/util/HudiTableTypeUtils.java	Adds centralized mapping of Hive input format class name -> HoodieTableType.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/util/HudiAvroSerializer.java	Adds Avro GenericRecord <-> Trino Block/Page serialization utilities.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/storage/TrinoHudiStorage.java	Routes inlinefs schemes to inline storage; changes getUri and setModificationTime behavior.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/storage/TrinoHudiInlineStorage.java	Implements HoodieStorage wrapper for inlinefs reads by sub-reading outer files.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/io/InlineSeekableDataInputStream.java	Adds a seekable view into a byte range of an underlying TrinoInputStream.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/file/HudiLogFile.java	Adds Jackson-serializable log file abstraction with split offsets.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/file/HudiFile.java	Introduces shared file abstraction interface.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/file/HudiBaseFile.java	Adds Jackson-serializable base file abstraction with split offsets.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/TimelineTable.java	Updates meta client creation to include table name context.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/HudiUtil.java	Adds schema helpers + cached schema field lookup; improves meta client error reporting surface.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/HudiTransactionManager.java	Tightens visibility of memoized metadata getter.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/HudiSplitSource.java	Updates meta client creation to include table name context.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/HudiPageSourceProvider.java	Ensures close() IOException is suppressed on the original exception.
plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/HudiErrorCode.java	Adds new error codes for unsupported table type, meta client, and schema errors.
plugin/trino-hudi/pom.xml	Adds dependencies for cache + Kryo and adjusts dependency plugin configuration.

Comments suppressed due to low confidence (1)

plugin/trino-hudi/src/test/java/io/trino/plugin/hudi/util/TestTupleDomainUtilsTest.java:31

The test class/file name TestTupleDomainUtilsTest is likely unintended duplication ("Test" suffix twice). This can make test discovery/reporting harder to read. Consider renaming to TestTupleDomainUtils (or another name matching the specific behavior under test).

class TestTupleDomainUtilsTest
{

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

Copilot · 2026-03-04T05:12:28Z

plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/io/InlineSeekableDataInputStream.java

+    public InlineSeekableDataInputStream(TrinoInputStream stream, long startOffset, long length)
+            throws IOException
+    {
+        super(stream);
+        this.startOffset = startOffset;
+        this.length = length;
+        stream.seek(startOffset);
+    }
+
+    @Override
+    public long getPos()
+            throws IOException
+    {
+        return super.getPos() - startOffset;
+    }
+
+    @Override
+    public void seek(long pos)
+            throws IOException
+    {
+        if (pos > length) {
+            throw new IOException(String.format(
+                    "Attempting to seek past inline content: position to seek to is %s but the length is %s",
+                    pos, length));
+        }
+        super.seek(startOffset + pos);
+    }


InlineSeekableDataInputStream enforces bounds for seek(), but it does not enforce the [startOffset, startOffset+length] boundary for reads. As written, a consumer can read past the inline segment into the outer file, which can corrupt inline block decoding. Consider tracking remaining bytes and overriding read()/read(byte[],..)/skip() to stop at EOF when pos == length (and optionally validate startOffset/length >= 0).

Copilot · 2026-03-04T05:12:28Z

plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/storage/TrinoHudiStorage.java

+        Location sameLocation = convertToLocation(path);
+        fileSystem.renameFile(sameLocation, sameLocation);


TrinoHudiStorage.setModificationTime ignores the provided modificationTimeInMillisEpoch parameter and attempts renameFile(path, path) instead. This does not actually set the requested timestamp and may be a no-op or even fail on some filesystems. If TrinoFileSystem cannot set mtime, it would be safer to keep this unsupported (throw) or implement semantics that match the API contract (e.g., via a filesystem-specific setLastModified if available).

Suggested change

Location sameLocation = convertToLocation(path);

fileSystem.renameFile(sameLocation, sameLocation);

throw new UnsupportedOperationException("TrinoHudiStorage does not support setting modification time");

Copilot · 2026-03-04T05:12:28Z

plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/util/TupleDomainUtils.java

+    public static boolean areAllFieldsReferenced(TupleDomain<String> tupleDomain, List<String> sourceFields)
+    {
+        Set<String> referenceColSet = new HashSet<>(TupleDomainUtils.getReferencedColumns(tupleDomain));
+        Set<String> sourceFieldSet = new HashSet<>(sourceFields);
+
+        return referenceColSet.containsAll(sourceFieldSet);
+    }
+
+    /**
+     * Check if at least one of the provided source field is referenced in the tupleDomain predicates.
+     */
+    public static boolean areSomeFieldsReferenced(TupleDomain<String> tupleDomain, List<String> sourceFields)
+    {
+        Set<String> referenceColSet = new HashSet<>(TupleDomainUtils.getReferencedColumns(tupleDomain));
+        for (String sourceField : sourceFields) {
+            if (referenceColSet.contains(sourceField)) {
+                return true;
+            }
+        }
+        return false;


TupleDomainUtils.areAllFieldsReferenced and areSomeFieldsReferenced will throw NullPointerException if sourceFields is null, while areDomainsInOrEqualOnly explicitly treats null sourceFields as false. Consider either adding a null check (returning false or true depending on intended semantics) or enforcing non-null via requireNonNull and documenting it, for consistency across this utility class.

Copilot · 2026-03-04T05:12:29Z

plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/storage/TrinoHudiInlineStorage.java

+    @Override
+    public URI getUri()
+    {
+        return URI.create(getScheme());


TrinoHudiInlineStorage.getUri() returns URI.create("inlinefs"), which results in a relative URI with scheme==null. If HoodieStorage#getUri is expected to carry the scheme, consider returning something like "inlinefs:///" (or otherwise ensuring URI.getScheme() returns "inlinefs").

Suggested change

return URI.create(getScheme());

// Ensure this returns an absolute URI whose scheme matches getScheme()

return URI.create(getScheme() + ":///");

Copilot · 2026-03-04T05:12:29Z

plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/util/TupleDomainUtils.java

+        // All recordKeys must have a domain else, return empty list (applicable to complexKeys)
+        // If a one of the recordKey in the set of complexKeys does not have a domain, we are unable to construct
+        // a complete complexKey
+        if (!recordKeyDomains.getDomains().get().keySet().containsAll(recordKeyFields)) {
+            return Collections.emptyList();
+        }
+
+        // Extract the domain mappings from the tuple domain
+        Map<String, Domain> domains = recordKeyDomains.getDomains().get();
+


constructRecordKeys calls recordKeyDomains.getDomains().get() without handling recordKeyDomains.isNone() (or domains Optional empty). If recordKeyDomains is TupleDomain.none(), this will throw NoSuchElementException. Consider explicitly checking isNone()/getDomains().isEmpty() early and returning an empty list (or otherwise handling the contradiction case).

Suggested change

// All recordKeys must have a domain else, return empty list (applicable to complexKeys)

// If a one of the recordKey in the set of complexKeys does not have a domain, we are unable to construct

// a complete complexKey

if (!recordKeyDomains.getDomains().get().keySet().containsAll(recordKeyFields)) {

return Collections.emptyList();

}

// Extract the domain mappings from the tuple domain

Map<String, Domain> domains = recordKeyDomains.getDomains().get();

// If the tuple domain is contradictory or has no domains, we cannot construct any record keys

if (recordKeyDomains.isNone()) {

return Collections.emptyList();

}

Optional<Map<String, Domain>> domainsOptional = recordKeyDomains.getDomains();

if (domainsOptional.isEmpty() || domainsOptional.get().isEmpty()) {

return Collections.emptyList();

}

Map<String, Domain> domains = domainsOptional.get();

// All recordKeys must have a domain else, return empty list (applicable to complexKeys)

// If one of the recordKeys in the set of complexKeys does not have a domain, we are unable to construct

// a complete complexKey

if (!domains.keySet().containsAll(recordKeyFields)) {

return Collections.emptyList();

}

Copilot · 2026-03-04T05:12:29Z

plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/util/TupleDomainUtils.java

+        if (regularColumnPredicates.isNone() || !regularColumnPredicates.getDomains().isPresent()) {
+            return true;
+        }
+        for (String regularColumn : regularColumns) {
+            Domain columnPredicate = regularColumnPredicates.getDomains().get().get(regularColumn);
+            Optional<Domain> currentColumnStats = Optional.ofNullable(domainsWithStats.get(regularColumn));
+            if (currentColumnStats.isEmpty()) {
+                // No stats for column
+            }
+            else {
+                Domain domain = currentColumnStats.get();
+                if (columnPredicate.intersect(domain).isNone()) {
+                    return false;
+                }


evaluateStatisticPredicate's early return treats regularColumnPredicates.isNone() as "may match" and returns true. Per the method contract, TupleDomain.none() means no row can satisfy the predicate, so the method should return false (skip file) in that case. Consider changing the early return to if (regularColumnPredicates.isNone()) { return false; } and handling the "no domains / all" cases separately.

Suggested change

if (regularColumnPredicates.isNone() || !regularColumnPredicates.getDomains().isPresent()) {

return true;

}

for (String regularColumn : regularColumns) {

Domain columnPredicate = regularColumnPredicates.getDomains().get().get(regularColumn);

Optional<Domain> currentColumnStats = Optional.ofNullable(domainsWithStats.get(regularColumn));

if (currentColumnStats.isEmpty()) {

// No stats for column

}

else {

Domain domain = currentColumnStats.get();

if (columnPredicate.intersect(domain).isNone()) {

return false;

}

// A 'None' TupleDomain means no row can satisfy the predicate; skip the file.

if (regularColumnPredicates.isNone()) {

return false;

}

Optional<Map<String, Domain>> predicateDomains = regularColumnPredicates.getDomains();

// An 'All' TupleDomain has no constraints; the file may match.

if (predicateDomains.isEmpty()) {

return true;

}

Map<String, Domain> predicateDomainsMap = predicateDomains.get();

for (String regularColumn : regularColumns) {

Domain columnPredicate = predicateDomainsMap.get(regularColumn);

if (columnPredicate == null) {

// No predicate for this column

continue;

}

Optional<Domain> currentColumnStats = Optional.ofNullable(domainsWithStats.get(regularColumn));

if (currentColumnStats.isEmpty()) {

// No stats for column

continue;

}

Domain domain = currentColumnStats.get();

if (columnPredicate.intersect(domain).isNone()) {

return false;

Copilot · 2026-03-04T05:12:30Z

plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/util/HudiAvroSerializer.java

+    public HudiAvroSerializer(List<HiveColumnHandle> columnHandles)
+    {
+        this.columnHandles = columnHandles;
+        this.columnTypes = columnHandles.stream().map(HiveColumnHandle::getType).toList();
+        // Fetches projected schema for #serialize() (source page required for Avro record conversion).
+        // buildRecordInPage() uses record#getSchema() directly, so schema is not needed there.
+        // Guard against empty column list (e.g. when only partition/synthesized columns are selected) since constructSchema fails on an empty column list.
+        var nonHiddenNames = columnHandles.stream().filter(ch -> !ch.isHidden()).map(HiveColumnHandle::getName).toList();
+        var nonHiddenTypes = columnHandles.stream().filter(ch -> !ch.isHidden()).map(HiveColumnHandle::getHiveType).toList();
+        this.schema = nonHiddenNames.isEmpty() ? null : constructSchema(nonHiddenNames, nonHiddenTypes);
+    }
+
+    public IndexedRecord serialize(SourcePage sourcePage, int position)
+    {
+        IndexedRecord record = new GenericData.Record(schema);
+        for (int i = 0; i < columnTypes.size(); i++) {
+            Object value = getValue(sourcePage, i, position);
+            record.put(i, value);
+        }
+        return record;


HudiAvroSerializer can construct schema=null when all provided HiveColumnHandles are hidden, but serialize() unconditionally does new GenericData.Record(schema) which will NPE. Additionally, schema is built only from non-hidden handles while serialize() iterates over all columnTypes and uses positional record.put(i, ...), which can mismatch the Avro schema field count/indexing if hidden columns are present. Consider either (a) rejecting hidden-only inputs early, (b) building a schema aligned with the positions you write, or (c) maintaining a mapping from channel -> schema field index and only serializing non-hidden columns.

Addressed, I added channelToSchemaFieldIndex[], which is computed once in the constructor.

Each non-hidden channel gets its sequential schema field index; hidden channels get -1. serialize() uses fieldIdx = channelToSchemaFieldIndex[i] for record.put(...), so the Avro field positions always match the schema regardless of where hidden columns appear in the channel list.

Below is the logic to ignore -1/hidden fields, so we should be safe now .

if (fieldIdx < 0) { continue; // hidden column — not part of the schema }

Copilot · 2026-03-04T05:12:30Z

plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/storage/TrinoHudiStorage.java

    public URI getUri()
    {
-        return URI.create("");
+        return URI.create(getScheme());


TrinoHudiStorage.getUri() currently returns URI.create(getScheme()) (e.g., "file"), which produces a relative URI with no scheme component (scheme==null, path=="file"). HoodieStorage#getUri is typically expected to include a scheme (e.g., "file:///") so that downstream URI-based logic can reliably extract the scheme. Consider returning URI.create(getScheme() + ":///") (or otherwise constructing a URI with the scheme set).

Suggested change

return URI.create(getScheme());

return URI.create(getScheme() + ":///");

Copilot · 2026-03-04T05:12:30Z

plugin/trino-hudi/src/main/java/io/trino/plugin/hudi/HudiUtil.java

+                    "Location of table %s does not contain Hudi table metadata: %s".formatted(tableName, basePath));
+        }
+        catch (Throwable e) {
+            throw new TrinoException(HUDI_META_CLIENT_ERROR,
+                    "Unable to load Hudi meta client for table %s (%s)".formatted(tableName, basePath));


buildTableMetaClient swallows the underlying exception details: both catch blocks drop the original exception as the cause, and the generic catch (Throwable) will also intercept Errors/interruptions. Consider catching a narrower exception type (e.g., RuntimeException) and passing the original exception as the cause when constructing TrinoException so failures are diagnosable.

Suggested change

"Location of table %s does not contain Hudi table metadata: %s".formatted(tableName, basePath));

}

catch (Throwable e) {

throw new TrinoException(HUDI_META_CLIENT_ERROR,

"Unable to load Hudi meta client for table %s (%s)".formatted(tableName, basePath));

"Location of table %s does not contain Hudi table metadata: %s".formatted(tableName, basePath),

e);

}

catch (RuntimeException e) {

throw new TrinoException(HUDI_META_CLIENT_ERROR,

"Unable to load Hudi meta client for table %s (%s)".formatted(tableName, basePath),

e);

Addressed, exception is added.

cla-bot · 2026-03-04T10:34:39Z

Thank you for your pull request and welcome to the Trino community. We require contributors to sign our Contributor License Agreement, and we don't seem to have you on file. Continue to work with us on the review and improvements in this PR, and submit the signed CLA to cla@trino.io. Photos, scans, or digitally-signed PDF files are all suitable. Processing may take a few days. The CLA needs to be on file before we merge your changes. For more information, see https://github.com/trinodb/cla

- HudiFile/HudiBaseFile/HudiLogFile: Trino-native wrappers for Hudi base files and log files - TupleDomainUtils: predicate helpers for domain-based index lookups - HudiAvroSerializer: bidirectional Avro <-> Trino type conversion - HudiTableTypeUtils: COW/MOR input format detection - InlineSeekableDataInputStream / TrinoHudiInlineStorage: support for reading log files embedded via InLineFS URI scheme

cla-bot · 2026-03-05T05:39:57Z

Thank you for your pull request and welcome to the Trino community. We require contributors to sign our Contributor License Agreement, and we don't seem to have you on file. Continue to work with us on the review and improvements in this PR, and submit the signed CLA to cla@trino.io. Photos, scans, or digitally-signed PDF files are all suitable. Processing may take a few days. The CLA needs to be on file before we merge your changes. For more information, see https://github.com/trinodb/cla

github-actions bot added the hudi Hudi connector label Mar 4, 2026

voonhous force-pushed the oss-master-upstream-1 branch from 7e963d1 to b1f8af0 Compare March 4, 2026 03:52

voonhous force-pushed the oss-master-upstream-1 branch from b1f8af0 to 8f2a118 Compare March 4, 2026 04:03

voonhous force-pushed the oss-master-upstream-1 branch from 8f2a118 to df98011 Compare March 4, 2026 04:08

voonhous force-pushed the oss-master-upstream-1 branch from df98011 to 71de823 Compare March 4, 2026 04:44

ebyhr requested a review from Copilot March 4, 2026 05:08

Copilot started reviewing on behalf of ebyhr March 4, 2026 05:08 View session

Copilot AI reviewed Mar 4, 2026

View reviewed changes

voonhous force-pushed the oss-master-upstream-1 branch from 6b58549 to 097226d Compare March 4, 2026 10:34

voonhous mentioned this pull request Mar 4, 2026

Add configuration, statistics infrastructure, index framework, and MOR read pipeline for Hudi connector #28533

Open

voonhous force-pushed the oss-master-upstream-1 branch from 097226d to 78209f8 Compare March 5, 2026 05:39

		Location sameLocation = convertToLocation(path);
		fileSystem.renameFile(sameLocation, sameLocation);

	Location sameLocation = convertToLocation(path);
	fileSystem.renameFile(sameLocation, sameLocation);
	throw new UnsupportedOperationException("TrinoHudiStorage does not support setting modification time");

	return URI.create(getScheme());
	// Ensure this returns an absolute URI whose scheme matches getScheme()
	return URI.create(getScheme() + ":///");

Conversation

voonhous commented Mar 4, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Additional context and related issues

Release notes

Uh oh!

cla-bot bot commented Mar 4, 2026

Uh oh!

cla-bot bot commented Mar 4, 2026

Uh oh!

cla-bot bot commented Mar 4, 2026

Uh oh!

cla-bot bot commented Mar 4, 2026

Uh oh!

cla-bot bot commented Mar 4, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Mar 4, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 4, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 4, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 4, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 4, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 4, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 4, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 4, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 4, 2026

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

cla-bot bot commented Mar 4, 2026

Uh oh!

cla-bot bot commented Mar 5, 2026

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

2 participants

voonhous commented Mar 4, 2026 •

edited

Loading