Apache
/
hbase
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

HBASE-11262 Avoid empty columns while doing bulk-load (Ashish Kumar)

This commit is contained in:
tedyu 2016-02-05 09:00:04 -08:00
parent 6f6a8ed71f
commit 64bac770d4
3 changed files with 21 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/ImportTsv.java
View File

@ -91,6 +91,7 @@ public class ImportTsv extends Configured implements Tool {
// If true, bad lines are logged to stderr. Default: false.
public final static String LOG_BAD_LINES_CONF_KEY = "importtsv.log.bad.lines";
public final static String SKIP_LINES_CONF_KEY = "importtsv.skip.bad.lines";
public final static String SKIP_EMPTY_COLUMNS = "importtsv.skip.empty.columns";
public final static String COLUMNS_CONF_KEY = "importtsv.columns";
public final static String SEPARATOR_CONF_KEY = "importtsv.separator";
public final static String ATTRIBUTE_SEPERATOR_CONF_KEY = "attributes.seperator";
@ -685,6 +686,7 @@ public class ImportTsv extends Configured implements Tool {
" table. If table does not exist, it is created but deleted in the end.\n" +
" -D" + SKIP_LINES_CONF_KEY + "=false - fail if encountering an invalid line\n" +
" -D" + LOG_BAD_LINES_CONF_KEY + "=true - logs invalid lines to stderr\n" +
" -D" + SKIP_EMPTY_COLUMNS + "=false - If true then skip empty columns in bulk import\n" +
" '-D" + SEPARATOR_CONF_KEY + "=|' - eg separate on pipes instead of tabs\n" +
" -D" + TIMESTAMP_CONF_KEY + "=currentTimeAsLong - use the specified timestamp for the import\n" +
" -D" + MAPPER_CONF_KEY + "=my.Mapper - A user-defined Mapper to use instead of " +

7
hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/TsvImporterMapper.java
View File

@ -58,6 +58,8 @@ extends Mapper<LongWritable, Text, ImmutableBytesWritable, Put>
/** Should skip bad lines */
private boolean skipBadLines;
/** Should skip empty columns*/
private boolean skipEmptyColumns;
private Counter badLineCount;
private boolean logBadLines;
@ -133,6 +135,8 @@ extends Mapper<LongWritable, Text, ImmutableBytesWritable, Put>
// configuration.
ts = conf.getLong(ImportTsv.TIMESTAMP_CONF_KEY, 0);
skipEmptyColumns = context.getConfiguration().getBoolean(
ImportTsv.SKIP_EMPTY_COLUMNS, false);
skipBadLines = context.getConfiguration().getBoolean(
ImportTsv.SKIP_LINES_CONF_KEY, true);
badLineCount = context.getCounter("ImportTsv", "Bad Lines");
@ -178,7 +182,8 @@ extends Mapper<LongWritable, Text, ImmutableBytesWritable, Put>
for (int i = 0; i < parsed.getColumnCount(); i++) {
if (i == parser.getRowKeyColumnIndex() || i == parser.getTimestampKeyColumnIndex()
|| i == parser.getAttributesKeyColumnIndex() || i == parser.getCellVisibilityColumnIndex()
|| i == parser.getCellTTLColumnIndex()) {
|| i == parser.getCellTTLColumnIndex() || (skipEmptyColumns
&& parsed.getColumnLength(i) == 0)) {
continue;
}
populatePut(lineBytes, parsed, put, i);

13
hbase-server/src/test/java/org/apache/hadoop/hbase/mapreduce/TestImportTsv.java
View File

@ -363,6 +363,19 @@ public class TestImportTsv implements Configurable {
doMROnTableTest(util, tn, FAMILY, data, args, 1, 4);
util.deleteTable(tn);
}
@Test
public void testSkipEmptyColumns() throws Exception {
Path bulkOutputPath = new Path(util.getDataTestDirOnTestFS(tn.getNameAsString()), "hfiles");
args.put(ImportTsv.BULK_OUTPUT_CONF_KEY, bulkOutputPath.toString());
args.put(ImportTsv.COLUMNS_CONF_KEY, "HBASE_ROW_KEY,HBASE_TS_KEY,FAM:A,FAM:B");
args.put(ImportTsv.SEPARATOR_CONF_KEY, ",");
args.put(ImportTsv.SKIP_EMPTY_COLUMNS, "true");
// 2 Rows of data as input. Both rows are valid and only 3 columns are no-empty among 4
String data = "KEY,1234,VALUE1,VALUE2\nKEY,1235,,VALUE2\n";
doMROnTableTest(util, tn, FAMILY, data, args, 1, 3);
util.deleteTable(tn);
}
private Tool doMROnTableTest(String data, int valueMultiplier) throws Exception {
return doMROnTableTest(util, tn, FAMILY, data, args, valueMultiplier,-1);