Limiting columns per record in CQL

问题

I've a problem which has been bothering me from quite while now. I'm scaling it down for simplification.

I've a column family in Cassandra defined as:

  CREATE TABLE "Test" (
  key text,                         
  column1 text,                    
  value text,                        
  PRIMARY KEY (key, column1)       
)

If I run a query in CQL as:

select * from "Test" where key in ('12345','34567');

It gives me something like:

  key  | column1 | value
-----------------------+---
 12345 |     764 |    764
 12345 |     836 |    836
 12345 |  123723 | 123723
 12345 |  155863 | 155863

  key  | column1 | value
-----------------------+---
 34567 |  159144 | 159144
 34567 |  159869 | 159869
 34567 |  160705 | 160705

Now my question is how can I limit my results to 2 rows max per record. I tried to use the following but didn't work.

select FIRST 10 'a'..'z' from "Test" where key in ('12345','34567'); 

- Not available in latest CQL version.

select * from "Test" where key in ('12345','34567') limit 2;

-Only limits total number of rows, not per record

回答1:

There is no way to have this type of limit in CQL3. You have to run a separate query for each partition.

If query latency is not an issue for you, you always can install SparkSQL/Hive on top of your Cassandra database for complex analytical queries like the one in your original question. You even can cache the result of these queries.

来源：https://stackoverflow.com/questions/30568525/limiting-columns-per-record-in-cql