1
0

HUDI-148 Small File selection logic for MOR must skip fileIds selected for pending compaction correctly

This commit is contained in:
Balaji Varadarajan
2019-06-08 12:40:08 -07:00
committed by vinoth chandar
parent 8c9980f4f5
commit a1483f2c5f
11 changed files with 48 additions and 27 deletions

View File

@@ -104,13 +104,15 @@ public interface TableFileSystemView {
*/
Stream<FileSlice> getLatestUnCompactedFileSlices(String partitionPath);
/**
* Stream all the latest file slices in the given partition with precondition that
* commitTime(file) before maxCommitTime
* Stream all latest file slices in given partition with precondition that commitTime(file) before maxCommitTime
*
* @param partitionPath Partition path
* @param maxCommitTime Max Instant Time
* @param includeFileSlicesInPendingCompaction include file-slices that are in pending compaction
*/
Stream<FileSlice> getLatestFileSlicesBeforeOrOn(String partitionPath,
String maxCommitTime);
String maxCommitTime, boolean includeFileSlicesInPendingCompaction);
/**
* Stream all "merged" file-slices before on an instant time

View File

@@ -476,13 +476,19 @@ public abstract class AbstractTableFileSystemView implements SyncableFileSystemV
}
@Override
public final Stream<FileSlice> getLatestFileSlicesBeforeOrOn(String partitionStr, String maxCommitTime) {
public final Stream<FileSlice> getLatestFileSlicesBeforeOrOn(String partitionStr, String maxCommitTime,
boolean includeFileSlicesInPendingCompaction) {
try {
readLock.lock();
String partitionPath = formatPartitionKey(partitionStr);
ensurePartitionLoadedCorrectly(partitionPath);
return fetchLatestFileSlicesBeforeOrOn(partitionPath, maxCommitTime)
.map(fs -> filterDataFileAfterPendingCompaction(fs));
Stream<FileSlice> fileSliceStream =
fetchLatestFileSlicesBeforeOrOn(partitionPath, maxCommitTime);
if (includeFileSlicesInPendingCompaction) {
return fileSliceStream.map(fs -> filterDataFileAfterPendingCompaction(fs));
} else {
return fileSliceStream.filter(fs -> !isPendingCompactionScheduledForFileId(fs.getFileGroupId()));
}
} finally {
readLock.unlock();
}

View File

@@ -166,9 +166,10 @@ public class PriorityBasedFileSystemView implements SyncableFileSystemView, Seri
}
@Override
public Stream<FileSlice> getLatestFileSlicesBeforeOrOn(String partitionPath, String maxCommitTime) {
return execute(partitionPath, maxCommitTime, preferredView::getLatestFileSlicesBeforeOrOn,
secondaryView::getLatestFileSlicesBeforeOrOn);
public Stream<FileSlice> getLatestFileSlicesBeforeOrOn(String partitionPath, String maxCommitTime,
boolean includeFileSlicesInPendingCompaction) {
return execute(partitionPath, maxCommitTime, includeFileSlicesInPendingCompaction,
preferredView::getLatestFileSlicesBeforeOrOn, secondaryView::getLatestFileSlicesBeforeOrOn);
}
@Override

View File

@@ -108,6 +108,8 @@ public class RemoteHoodieTableFileSystemView implements SyncableFileSystemView,
public static final String LAST_INSTANT_TS = "lastinstantts";
public static final String TIMELINE_HASH = "timelinehash";
public static final String REFRESH_OFF = "refreshoff";
public static final String INCLUDE_FILES_IN_PENDING_COMPACTION_PARAM = "includependingcompaction";
private static Logger log = LogManager.getLogger(RemoteHoodieTableFileSystemView.class);
@@ -327,8 +329,11 @@ public class RemoteHoodieTableFileSystemView implements SyncableFileSystemView,
}
@Override
public Stream<FileSlice> getLatestFileSlicesBeforeOrOn(String partitionPath, String maxCommitTime) {
Map<String, String> paramsMap = getParamsWithAdditionalParam(partitionPath, MAX_INSTANT_PARAM, maxCommitTime);
public Stream<FileSlice> getLatestFileSlicesBeforeOrOn(String partitionPath, String maxCommitTime,
boolean includeFileSlicesInPendingCompaction) {
Map<String, String> paramsMap = getParamsWithAdditionalParams(partitionPath,
new String[]{MAX_INSTANT_PARAM, INCLUDE_FILES_IN_PENDING_COMPACTION_PARAM},
new String[]{maxCommitTime, String.valueOf(includeFileSlicesInPendingCompaction)});
try {
List<FileSliceDTO> dataFiles = executeRequest(LATEST_SLICES_BEFORE_ON_INSTANT_URL, paramsMap,
new TypeReference<List<FileSliceDTO>>() {

View File

@@ -351,7 +351,7 @@ public class HoodieTableFileSystemViewTest {
assertEquals("Log File Order check", fileName2, logFiles.get(2).getFileName());
assertEquals("Log File Order check", fileName1, logFiles.get(3).getFileName());
fileSliceList = rtView.getLatestFileSlicesBeforeOrOn(partitionPath, deltaInstantTime5)
fileSliceList = rtView.getLatestFileSlicesBeforeOrOn(partitionPath, deltaInstantTime5, true)
.collect(Collectors.toList());
assertEquals("Expect only one file-id", 1, fileSliceList.size());
fileSlice = fileSliceList.get(0);
@@ -672,7 +672,7 @@ public class HoodieTableFileSystemViewTest {
assertTrue(filenames.contains(FSUtils.makeDataFileName(commitTime4, TEST_WRITE_TOKEN, fileId3)));
filenames = Sets.newHashSet();
List<HoodieLogFile> logFilesList = rtView.getLatestFileSlicesBeforeOrOn("2016/05/01", commitTime4)
List<HoodieLogFile> logFilesList = rtView.getLatestFileSlicesBeforeOrOn("2016/05/01", commitTime4, true)
.map(slice -> slice.getLogFiles()).flatMap(logFileList -> logFileList)
.collect(Collectors.toList());
assertEquals(logFilesList.size(), 4);
@@ -706,7 +706,7 @@ public class HoodieTableFileSystemViewTest {
}
logFilesList =
rtView.getLatestFileSlicesBeforeOrOn("2016/05/01", commitTime3).map(slice -> slice.getLogFiles())
rtView.getLatestFileSlicesBeforeOrOn("2016/05/01", commitTime3, true).map(slice -> slice.getLogFiles())
.flatMap(logFileList -> logFileList).collect(Collectors.toList());
assertEquals(logFilesList.size(), 1);
assertTrue(logFilesList.get(0).getFileName()
@@ -1135,7 +1135,7 @@ public class HoodieTableFileSystemViewTest {
assertEquals("Log File Order check", fileName3, logFiles.get(1).getFileName());
assertEquals("Log File Order check", fileName1, logFiles.get(2).getFileName());
fileSliceList = rtView.getLatestFileSlicesBeforeOrOn(partitionPath, deltaInstantTime5)
fileSliceList = rtView.getLatestFileSlicesBeforeOrOn(partitionPath, deltaInstantTime5, true)
.collect(Collectors.toList());
assertEquals("Expect only one file-id", 1, fileSliceList.size());
fileSlice = fileSliceList.get(0);
@@ -1147,6 +1147,11 @@ public class HoodieTableFileSystemViewTest {
assertEquals("Log files must include only those after compaction request", 2, logFiles.size());
assertEquals("Log File Order check", fileName4, logFiles.get(0).getFileName());
assertEquals("Log File Order check", fileName3, logFiles.get(1).getFileName());
// Check getLatestFileSlicesBeforeOrOn excluding fileIds in pending compaction
fileSliceList = rtView.getLatestFileSlicesBeforeOrOn(partitionPath, deltaInstantTime5, false)
.collect(Collectors.toList());
assertEquals("Expect empty list as file-id is in pending compaction", 0, fileSliceList.size());
});
Assert.assertEquals(3, fsView.getPendingCompactionOperations().count());