Details
Major Description
In the Thrift 1 definition file it says:
/**
* Get a scanner on the current table starting at the specified row and
* ending at the last row in the table. Return the specified columns.
*
* @return scanner id to be used with other scanner procedures
*/
ScannerID scannerOpen(
/** name of table */
1:Text tableName,
/**
* Starting row in table to scan.
* Send "" (empty string) to start at the first row.
*/
2:Text startRow,
/**
* columns to scan. If column name is a column family, all
* columns of the specified column family are returned. It's also possible
* to pass a regex in the column qualifier.
*/
3:list<Text> columns,
/** Scan attributes */
4:map<Text, Text> attributes
) throws (1:IOError io)
Especially that "columns" can contain a regex for the qualifier. The code though does not support that currently:
@Override
public int scannerOpen(ByteBuffer tableName, ByteBuffer startRow,
List<ByteBuffer> columns,
Map<ByteBuffer, ByteBuffer> attributes) throws IOError {
try {
HTable table = getTable(tableName);
Scan scan = new Scan(getBytes(startRow));
addAttributes(scan, attributes);
if(columns != null && columns.size() != 0) {
for(ByteBuffer column : columns) {
byte [][] famQf = KeyValue.parseColumn(getBytes(column));
if(famQf.length == 1) {
scan.addFamily(famQf[0]);
} else {
scan.addColumn(famQf[0], famQf[1]);
}
}
}
return addScanner(table.getScanner(scan));
} catch (IOException e) {
LOG.warn(e.getMessage(), e);
throw new IOError(e.getMessage());
}
}
It parses the columns as literals, and sets up the Scan without supporting the appropriate QualifierFilter (or others) with RegexStringComparator.