Order matters when adding columns to a scan, using some of the add column
methods.
----------------------------------------------------------------------------------
Key: HBASE-2071
URL: https://issues.apache.org/jira/browse/HBASE-2071
Project: Hadoop HBase
Issue Type: Bug
Components: client
Affects Versions: 0.20.2
Environment: Calling the Java API, using Clojure 1.1 alpha. Client
machine is a Mac OSX 10.6, using HBase server in non-distributed mode (locally).
Reporter: Kyle Oba
Priority: Minor
When adding columns to a Scan, using "addColumns" or "addColumn" (with single
argument) the order in which they are added matters. When adding columns as
distinct family and qualifier (2 arguments) this is not a problem. It is only
a problem when adding as a single argument.
For example:
user> (scan "t1" {:columns [ "f1:" "f1:hello" ] })
cols are: [f1: f1:hello]
col: f1: --
col: f1:hello -- world
user> (scan "t1" {:columns [ "f1:hello" "f1:" ] })
cols are: [f1:hello f1:]
col: f1:hello -- world
col: f1: -- v1
In the first call to "scan," the arguments are in lexicographic order, and this
results in the value associated with "f1" to be left out. In the second call
to "scan," the arguments are in reverse order, and the results are valid.
Sorry the example is in Clojure. In essence, this is what's happending:
This doesn't work:
myScan.addColumn("f1:")
myScan.addColumn("f1:hello")
This works:
myScan.addColumn("f1:hello")
myScan.addColumn("f1:")
I have omitted the code which creates a scanner with this scan, returns the
results, and then prints them.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
