On Tue, Jan 12, 2010 at 2:01 PM, Something Something < [email protected]> wrote:
> All the Keys returned by this code are guaranteed to be in order by the key > values, correct? > No. Explicitly, the order is NOT guaranteed, for performance reasons (I don't see this in the javadoc. It should be there). If just getting latest version from a set of columns, I believe it will come across in the right order but not guaranteed. To get a guaranteed sorted order, call http://hadoop.apache.org/hbase/docs/current/api/org/apache/hadoop/hbase/client/Result.html#sorted()instead of the raw() method. The sort will be done client-side. > > Also, for some other table I am retrieving all column names for a > particular > key, and those all seem to be in the correct order as well. Is this always > guaranteed? > Call sort if you need guarantee. St.Ack > > On Tue, Jan 12, 2010 at 11:12 AM, stack <[email protected]> wrote: > > > Setup the scanner and next it as you did previous. Then on the Result > > object, do something like: > > > > for (KeyValue kv: result.raw()) { > > System.out.println(Bytes.toString(kv.getRow()) + " " + > > Bytes.toString(kv.getValue())); > > } > > > > St.Ack > > > > On Tue, Jan 12, 2010 at 10:24 AM, Something Something < > > [email protected]> wrote: > > > > > Sorry. That was a typo. In any case, it seems like I am using the > wrong > > > API. > > > > > > Here's what my table contains: > > > > > > > > > ABC_111 column=info:estimate, > > timestamp=1263319888463, > > > value=179.59 > > > ABC_222 column=info:estimate, > > timestamp=1263319888463, > > > value=191.50 > > > ABC_333 column=info:estimate, > > timestamp=1263319888463, > > > value=180.65 > > > ABC_444 column=info:estimate, > > timestamp=1263319888463, > > > value=183.63 > > > & so on.... > > > > > > > > > I want to retrieve: > > > > > > ABC_111 179.59 > > > ABC_222 191.50 > > > ABC_333 180.65 > > > ABC_444 183.63 > > > & so on... > > > > > > What API should I use? Please let me know. Thanks for your help. > > > > > > > > > On Tue, Jan 12, 2010 at 9:36 AM, stack <[email protected]> wrote: > > > > > > > See below: > > > > > > > > On Tue, Jan 12, 2010 at 9:12 AM, Something Something < > > > > [email protected]> wrote: > > > > > > > > > > NavigableMap<byte[], NavigableMap<byte[], > > NavigableMap<Long, > > > > > byte[]>>> map = result.getMap(); > > > > > > > > > > > > > Above returns a map keyed by families: > > > > > > > > > > > > > > http://hadoop.apache.org/hbase/docs/current/api/org/apache/hadoop/hbase/client/Result.html#getMap%28%29 > > > > > > > > > > > > > > > > > > > > > for (Map.Entry<byte[], NavigableMap<byte[], > > > NavigableMap<Long, > > > > > byte[]>>> entry : map.entrySet()) { > > > > > byte[] key = entry.getKey(); > > > > > > > > > > > > > This is family name, not key. > > > > > > > > > > > > > > > > > *LOG.info("key = " + Bytes.toString(key));* > > > > > NavigableMap<byte[], NavigableMap<Long, byte[]>> value = > > > > > entry.getValue(); > > > > > > > > > > > > > This is a map keyed by column qualifiers. > > > > > > > > > > > > > > > > > for (Entry<byte[], NavigableMap<Long, byte[]>> entry1 > : > > > > > value.entrySet()) { > > > > > byte[] key1 = entry1.getKey(); > > > > > *LOG.info("key1 = " + Bytes.toString(key1));* > > > > > > > > > > > > > > > > > This is the family qualifier. > > > > > > > > > > > > > > > > > NavigableMap<byte[], NavigableMap<Long, byte[]>> > > > value1 > > > > = > > > > > entry.getValue(); > > > > > > > > > > > > > > > > > I do not think you intended to do this. I think you meant entry1, > not > > > > 'entry' and map type should be NavigableMap<Long, byte[]> rather than > > > > above. > > > > > > > > St.Ack > > > > > > > > > > > > > > > > > for (Entry<byte[], NavigableMap<Long, byte[]>> > > entry2 > > > : > > > > > value1.entrySet()) { > > > > > String key2 = Bytes.toString(entry2.getKey()); > > > > > *LOG.info("key2 = " + key2);* > > > > > > > > > > } > > > > > } > > > > > } > > > > > } > > > > > > > > > > > > > > >
