/**
* Builds a TableRecordReader. If no TableRecordReader was provided, uses the default.
*
* @see org.apache.hadoop.mapred.InputFormat#getRecordReader(InputSplit, JobConf, Reporter)
*/
public RecordReader<ImmutableBytesWritable, Result> getRecordReader(
InputSplit split, JobConf job, Reporter reporter) throws IOException {
TableSplit tSplit = (TableSplit) split;
TableRecordReader trr = this.tableRecordReader;
// if no table record reader was provided use default
if (trr == null) {
trr = new TableRecordReader();
}
trr.setStartRow(tSplit.getStartRow());
trr.setEndRow(tSplit.getEndRow());
trr.setHTable(this.table);
trr.setInputColumns(this.inputColumns);
trr.setRowFilter(this.rowFilter);
trr.init();
return trr;
}
public TableJoinRecordReader(
JobConf jobConf, CloudataConf conf, TableSplit tableSplit, Reporter reporter)
throws IOException {
this.conf = conf;
String mergeEvaluatorClass = tableSplit.getInputTableInfo().getMergeEvaluatorClass();
MergeEvaluator mergeEvaluator = null;
if (mergeEvaluatorClass != null && mergeEvaluatorClass.length() > 0) {
try {
mergeEvaluator = (MergeEvaluator) Class.forName(mergeEvaluatorClass).newInstance();
} catch (Exception e) {
LOG.error("mergeEvaluator:" + mergeEvaluatorClass + "," + e.getMessage());
IOException err = new IOException(e.getMessage() + ":" + mergeEvaluatorClass);
err.initCause(e);
throw err;
}
}
RowFilter splitRowFilter = tableSplit.getRowFilter();
InputTableInfo inputTableInfo = tableSplit.getInputTableInfo();
this.startRowKey = splitRowFilter.getStartRowKey();
this.endRowKey = splitRowFilter.getEndRowKey();
RowFilter rowFilter = inputTableInfo.getRowFilter();
rowFilter.setStartRowKey(startRowKey);
rowFilter.setEndRowKey(endRowKey);
CTable ctable = CTable.openTable(conf, inputTableInfo.getTableName());
TableScanner pivotScanner = null;
TableScanner targetScanner = null;
try {
pivotScanner =
ScannerFactory.openScanner(ctable, rowFilter, TableScanner.SCANNER_OPEN_TIMEOUT);
Row.Key firstRowKey = null;
try {
// 첫번째 Tablet이 아닌 경우에는 첫번째 row는 무시한다.
if (!startRowKey.equals(Row.Key.MIN_KEY)) {
Row pivotRow = pivotScanner.nextRow();
if (pivotRow == null) {
end = true;
return;
}
if (firstRowKey == null) {
firstRowKey = pivotRow.getKey();
}
if (firstRowKey.equals(pivotRow.getKey())) {
pivotRow = pivotScanner.nextRow();
if (pivotRow == null) {
end = true;
return;
}
}
pivotScanner.close();
rowFilter.setStartRowKey(firstRowKey);
pivotScanner =
ScannerFactory.openScanner(ctable, rowFilter, TableScanner.SCANNER_OPEN_TIMEOUT);
} else {
firstRowKey = startRowKey;
}
} catch (Exception e) {
if (pivotScanner != null) {
pivotScanner.close();
}
throw e;
}
RowFilter joinRowFilter = inputTableInfo.getJoinRowFilter();
if (mergeEvaluator != null) {
if (!firstRowKey.equals(Row.Key.MIN_KEY)) {
joinRowFilter.setStartRowKey(mergeEvaluator.parseTargetRowKey(firstRowKey, 0));
} else {
joinRowFilter.setStartRowKey(Row.Key.MIN_KEY);
}
if (!rowFilter.getEndRowKey().equals(Row.Key.MAX_KEY)) {
joinRowFilter.setEndRowKey(mergeEvaluator.parseTargetRowKey(rowFilter.getEndRowKey(), 0));
} else {
joinRowFilter.setEndRowKey(Row.Key.MAX_KEY);
}
} else {
joinRowFilter.setStartRowKey(firstRowKey);
joinRowFilter.setEndRowKey(rowFilter.getEndRowKey());
}
reporter.setStatus(
inputTableInfo.getTableName()
+ ":"
+ startRowKey
+ " ~ "
+ endRowKey
+ ", "
+ inputTableInfo.getJoinTableName()
+ ":"
+ joinRowFilter.getStartRowKey()
+ " ~ "
+ joinRowFilter.getEndRowKey());
// pivot table의 startRow, endRow에 대응하는 target table의 scanner 생성
CTable targetTable = CTable.openTable(conf, inputTableInfo.getJoinTableName());
targetScanner =
ScannerFactory.openScanner(targetTable, joinRowFilter, TableScanner.SCANNER_OPEN_TIMEOUT);
this.scanner =
new MergeScanner(new TableScanner[] {pivotScanner, targetScanner}, mergeEvaluator);
} catch (Exception e) {
if (targetScanner != null) {
targetScanner.close();
}
IOException err = new IOException(e.getMessage());
err.initCause(e);
throw err;
}
}
