How to count columns from multiple rows - cassandra

Question

How can I count number of columns in different rows of a column family?
I am a Cassandra newbie. I do not know a starting point. The only option I have is to make the application fetch data for each row at a time. It does not sound right to me. I am using Hector to connect to Cassandra.

Answer 1

this is how you will get total column count in particular rowkey
sliceQuery.setColumnFamily("**your column family**");
sliceQuery.setKey("**your row key**");
sliceQuery.setRange(null, null, false, Integer.MAX_VALUE);
QueryResult<ColumnSlice<String, String>> result = sliceQuery.execute();
ColumnSlice<String, String> cs = result.get();
long noOfColumnInRowKey=result.get().getColumns().size();

Answer 2

Assume that you have wide row (Lets create it using CLI)
create column family cf3
with column_type = 'Standard' and
comparator = 'TimeUUIDType' and
key_validation_class = 'UTF8Type' and
default_validation_class = 'UTF8Type';
This is what I see in CQL3:
cqlsh:ks> desc table cf3;
CREATE TABLE cf3 (
key text,
column1 timeuuid,
value text,
PRIMARY KEY (key, column1)
) WITH COMPACT STORAGE AND
bloom_filter_fp_chance=0.010000 AND
caching='KEYS_ONLY' AND
comment='' AND
dclocal_read_repair_chance=0.000000 AND
gc_grace_seconds=864000 AND
read_repair_chance=0.100000 AND
replicate_on_write='true' AND
populate_io_cache_on_flush='false' AND
compaction={'class': 'SizeTieredCompactionStrategy'} AND
compression={'sstable_compression': 'SnappyCompressor'};
I inserted some values from CQL3, which makes you feel like good ol' MySQL
cqlsh:ks> insert into cf3 (key, column1, value) values ('user1', now(), 'time5');
cqlsh:ks> select * from cf3;
key | column1 | value
-------+--------------------------------------+-------
user1 | f0c687b0-d114-11e2-8002-2f4261da0d90 | time1
user1 | fb9fa130-d114-11e2-8002-2f4261da0d90 | time2
user1 | 09512f10-d115-11e2-8002-2f4261da0d90 | time3
user1 | 0f5c93e0-d115-11e2-8002-2f4261da0d90 | time4
user1 | 21155220-d115-11e2-8002-2f4261da0d90 | time5
But it's your wide-row (as seen from CLI)
[default#ks] list cf3;
Using default limit of 100
Using default column limit of 100
-------------------
RowKey: user1
=> (column=f0c687b0-d114-11e2-8002-2f4261da0d90, value=time1, timestamp=1370789864363000)
=> (column=fb9fa130-d114-11e2-8002-2f4261da0d90, value=time2, timestamp=1370789882563000)
=> (column=09512f10-d115-11e2-8002-2f4261da0d90, value=time3, timestamp=1370789905537000)
=> (column=0f5c93e0-d115-11e2-8002-2f4261da0d90, value=time4, timestamp=1370789915678000)
=> (column=21155220-d115-11e2-8002-2f4261da0d90, value=time5, timestamp=1370789945410000)
1 Row Returned.
Elapsed time: 105 msec(s).
Now, you wanted to count number of columns starting from a given time onwards. Right? Here is CQL3 for that.
cqlsh:ks> select count(*) from cf3 where key = 'user1' and column1 >= 09512f10-d115-11e2-8002-2f4261da0d90 ;
count
-------
3
Now, I am somewhat doubtful what goes beneath. But, my intuition says that actually all the columns gets fetched at coordinator node and counted in memory. Which is probably somewhat similar to what you were planning to manually on client machine.
Also, I am unaware if cassandra-cli provides such functionality, but you mentioned you are using Hector. So, you can leverage get_count or CountQuery like mentioned here except have null as range finish and large count value. Like this:
CountQuery<String, String> cq = HFactory.createCountQuery(keyspace, StringSerializer.get(), TimeUUIDSerializer.get());
cq.setColumnFamily(cf).setKey("user1");
cq.setRange(timestamp, null, Integer.MAX_VALUE);
QueryResult<Integer> r = cq.execute();
(uncompiled code above)
HTH
Old answer:
See Hector documentation:
CQL:
CqlQuery<String,String,Long> cqlQuery = new CqlQuery<String,String,Long>(keyspace, se, se, le);
cqlQuery.setQuery("SELECT COUNT(*) FROM StandardLong1 WHERE KEY = 'cqlQueryTest_key1'");
QueryResult<CqlRows<String,String,Long>> result = cqlQuery.execute();
assertEquals(2, result.get().getAsCount());
You may just miss the WHERE condition and use LIMIT to get your purpose solved.