Hi Jilil,
I am new to HBase. I used filters on three different columns on a single table,
as shown below. The three filters are AND-ed together (must pass all). It
works well for me. Not sure it would help you or not.
FilterList filterList = new FilterList();
if (minClickCount != null) {
BinaryComparator bComparater = new
BinaryComparator(Bytes.toBytes(String.valueOf(minClickCount)));
SingleColumnValueFilter
filter = new
SingleColumnValueFilter(Bytes.toBytes(Constants.IMPRESSION_SUMMARY_COLUMN_FAMILY),
Bytes.toBytes(Constants.IMPRESSION_SUMMARY_CLICK_COUNT),
CompareFilter.CompareOp.GREATER_OR_EQUAL, bComparater);
filter.setFilterIfMissing(true);
filterList.addFilter(filter);
}
if (minCTR != null) {
BinaryComparator bComparater = new
BinaryComparator(Bytes.toBytes(String.valueOf(minCTR)));
SingleColumnValueFilter filter = new
SingleColumnValueFilter(Bytes.toBytes(Constants.IMPRESSION_SUMMARY_COLUMN_FAMILY),
Bytes.toBytes(Constants.IMPRESSION_SUMMARY_CLICK_THROUGH_RATE),
CompareFilter.CompareOp.GREATER_OR_EQUAL, bComparater);
filter.setFilterIfMissing(true);
filterList.addFilter(filter);
}
if (proID != null) {
SingleColumnValueFilter filter = new
SingleColumnValueFilter(Bytes.toBytes(Constants.IMPRESSION_SUMMARY_COLUMN_FAMILY),
Bytes.toBytes(Constants.IMPRESSION_SUMMARY_PRO_ID),
CompareFilter.CompareOp.EQUAL, Bytes.toBytes(proID));
filter.setFilterIfMissing(true);
filterList.addFilter(filter);
}
scan.setFilter(filterList);
Jack.
-----Original Message-----
From: [email protected] [mailto:[email protected]]
Sent: Wednesday, September 15, 2010 6:19 AM
To: [email protected]
Subject: Server side filters
Hello all,
I am trying to use HBase for a project. The data base will hold
billions of rows.
I am struggling to understand how filters work. In particular, I want
to be able to scan for rows that contain all of these columns for
example:
numbers:three & numbers:five & numbers:seven
I tried to set a FilterList with MustMatchAll operator and with a
QualifierFilter for each of the colum qualifiers I am looking for.
The problem is that the scanner returns no results if I set the
filters as above.
Is there a way to get only the rows that contain a set of column
qualifiers? server-side.
I know how to do it client side, by iterating through the results, but
I don't want my client to iterate through a billion rows. It wouldn't
make sense.
Many thanks,
Jalil
----- End forwarded message -----
