[HBASE-4811] Support reverse Scan

• Type: New Feature
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.20.6, 0.94.7
• Fix Version/s: 0.98.0
• Component/s: Client
• Hadoop Flags: Reviewed

Release Note

== What is it?

HBase 0.98 introduces a reverse scanner, so that you can scan rows in reverse order as well as the default order. Previously, if you wanted to scan in reverse, you needed to save rows in reverse order.

== Why do you want it?

One common use case is data stored with timestamps. Previously, when you created the schema, you needed to make a design decision about whether you wanted faster access to records with old timestamps or new ones. If you wanted to be able to scan in reverse, you needed to store the data twice, sorted forward and reverse.

With the reverse scanner in HBase 0.98, you can choose whether to scan your data in either direction. The reverse scanner is only a few percent slower than the default scanner.

== How do you set it up?
No setup is required.

== How do you use it?
Use the Scan.setReversed(boolean reversed) API call:
Scan.setReversed(true)

If you specify a startRow and stopRow, to scan in reverse, the startRow needs to be lexicographically after the stopRow.

Refer to the API documentation for more information about the Scan API (http://hbase.apache.org/apidocs/org/apache/hadoop/hbase/client/Scan.html).

NOTE: Rows are reversed only. Columns in a row still sort forward even if you specify reverse on your scan.

• Tags: Phoenix
• Assignee: Chunhui Shen
• Reporter: John Carrino
• Created: 17/Nov/11 19:03
• Updated: 21/Jun/17 07:50
• Resolved: 02/Dec/13 02:23

Description

Reversed scan means scan the rows backward.
And StartRow bigger than StopRow in a reversed scan.

For example, for the following rows:

aaa/c1:q1/value1
aaa/c1:q2/value2
bbb/c1:q1/value1
bbb/c1:q2/value2
ccc/c1:q1/value1
ccc/c1:q2/value2
ddd/c1:q1/value1
ddd/c1:q2/value2
eee/c1:q1/value1
eee/c1:q2/value2

you could do a reversed scan from 'ddd' to 'bbb'(exclude) like this:

Scan scan = new Scan();
scan.setStartRow('ddd');
scan.setStopRow('bbb');
scan.setReversed(true);
for(Result result:htable.getScanner(scan)){
System.out.println(result);
}

Aslo you could do the reversed scan with shell like this:
hbase> scan 'table',

{REVERSED => true,STARTROW=>'ddd', STOPROW=>'bbb'}

And the output is:
ddd/c1:q1/value1
ddd/c1:q2/value2
ccc/c1:q1/value1
ccc/c1:q2/value2

All the documentation I find about HBase says that if you want forward and reverse scans you should just build 2 tables and one be ascending and one descending. Is there a fundamental reason that HBase only supports forward Scan? It seems like a lot of extra space overhead and coding overhead (to keep them in sync) to support 2 tables.

I am assuming this has been discussed before, but I can't find the discussions anywhere about it or why it would be infeasible.

Issue Links

• incorporates: HBASE-11154 Document how to use Reverse Scan API
• is related to: HBASE-9999 Add support for small reverse scan
• relates to: HBASE-10016 Make use of backward seek implemented in reverse scan support for range queries.
• relates to: HBASE-10421 Investigate scan performance in 0.96 and later