We are exploring SPARK for cassandra in order to over come limitations with CQL.
We were initially restricted to CQL but faced few road blocks/hurdles over RDBMS. To name a few as below
For comparing >(Greater than) and < (Less than) on a column, we are restricted to have the columns in Clustering key. Even If I have a column in Clustering, I should still provide the Partition key to do < or > on clustering key.
Can't check for NULL on any column value
In order to query on any column other Partition key, we have to create index on that column
ORDER BY a column which isn't a CLUSTERING KEY
GROUP BY Limitations
Join Tables
I am a newbie with cassandra and end up in revisiting my schema often due to the limitations.
Hence similar to HIVE/PIG for HDFS, What additional benefits does Spark give over CQL ?
CQL is not a replacement for SQL. It is really designed for pulling out values from a few, usually one, partition key, and as you pointed out, does not do any sort of aggregation, grouping, very limited sorting, etc. (though Cassandra 3.0 will have UDFs and UDAs).
Here is what Spark offers over CQL:
General aggregation and querying via DataFrames and SQL, including JOINs, GROUP BY, ORDER BY, and UDFs
Significantly faster queries -- orders of magnitude faster -- if you cache the Cassandra data in memory using sqlContext.cacheTable
Integrated machine learning, statistics, graph processing, and virtually any kind of distributed computation you can imagine, using Scala, Java, Python, and R APIs
Ability to ETL in and out of Cassandra tables from and to many other data sources - including various HDFS formats, Amazon S3, DBMSes, Mongo, and most other databases today
Spark is really a completely different beast from CQL. It offers complex analytics over vast quantities of data, CQL doesn't. However, there are some limitations as well:
Spark is not good at highly concurrent queries. For that, you want to keep queries simple and use CQL to pull out a very small amount of data.
Caching data in Spark is not HA and cannot update as you write new data into C*
If you want very fast analytical queries over Cassandra with support for updates and no need to cache, then check out my project http://github.com/tuplejump/FiloDB.
Related
I understand that for Cassandra data is sorted on the cluster key only per partition key.
I am wondering if Cassandra has optimizations on global scans. Lets say that the cluster key is an integer value, if I want to search over all data on a Cassandra cluster to find collections with values < 3. The Cassandra query engine will not need to continue looking at collections in a partition after encountering a number >= 3. Are there APIs (such as CDK) offered by Cassandra which exercise these optimizations?
There isn't a native CQL optimisation available for full table scans -- they will always be bad since Cassandra is optimised for OLTP workloads.
There are however optimisations done by the spark-cassandra-connector for analytics (OLAP) workloads with Spark.
OLTP vs OLAP are worlds apart so you have to use the right tool for the job. Cheers!
Querying by partition key is the best way the query in Cassandra. If you want to use clustering key for querying then you can use "ALLOW FILTERING" option. But it is recommended not to use "ALLOW FILTERING" in production.
For scanning complete table and filtering some data, you can use spark to do your work. Why bother C* for which it is not designed, better to take help of its friends (spark in this case)
I am building a near real time/ microbatch data application with Cassandra as the lookup store. Each incremental run has ~40K records, while the Cassandra table has about 150 million records. In each run, I need to lookup the id field and get some attributes from Cassandra. These lookups can be random (not any time/ region/ country dependency), so there is no clear partitioning scheme.
How should I try to partition the Cassandra table to ensure decent/ good performance (for microbatches running every 15-30 mins)?
Apart from partitioning, any other tips?
joinWithCassandraTable and leftJoinWithCassandraTable functions were specifically designed for efficient data lookup in Cassandra from Spark jobs. It performs fetching of data by primary or partition key, and because it's executed by multiple executors in parallel, it could be fast (although ~40K could still take time, but it depends on size of your Cassandra and Spark clusters). See the SCC's documentation for detailed information how to use it - but remember, that these functions are available only in RDD API. The DataStax's version of connector has support for so-called "DirectJoin" - efficient joins with Cassandra in the DataFrame API.
Regarding partitioning - it depends on how do you perform lookup - you have 1 record in Cassandra matching one record in Spark? If yes, then just use this ID as primary key (it's equal to partition key in this case).
Not sure if the question's title sounds crazy but I thought about this and I'd like to check the "validity" of the pro's and con's I imagine.
The ideal C* query in "production" targets only one partition, possibly with additional restrictions on the clustering keys. A data model should be designed with that in mind.
However, for analytics jobs, e.g. using Spark, the queries would not work like that: "searching" for specific partitions is often needed (and I could not find a way to do that properly with SparkSQL and the dataframe API) and it should not work like this: a Spark job should target many partitions to spread over all the co-located Spark/Cassandra nodes.
My data model works in such a way that acquiring my data in real time inserts partitions as a whole. My partitions are "atomic": a large analytics job with Spark will mainly correlate data within one partition (which is good as it allows data locality for the Spark executor) but my main problem is to find on which partitions I want to operate.
So, what about duplicating my partition key and have it as a clustering key as well? This would allow me to build a SASI index on it and have the "best of both worlds" just at the cost of the additional storage.
Would this be a sound strategy?
As we read about Cassandra, before, we decided to choose it as our main database.The most important, useful and special feature which encourage us to choose this db, was Scalability, which helps us using Large volumes of data.
But, It seems that, it can not meet our requirements by itself. I asked some questions about our requirement in Stackoverfolw and how we can response them using Cassandra, and the answer was using alternative tools on top of Cassandra as Spark, Solr, DSE Search Tools and etc.
Our case is BIG Data Really, but we will have a large variety of Queries, too.
With these explanations, is it wise to stay on Cassandra?... Or It's better to switch to another db?
Cassandra is not adequate for ad-hoc queries, so I would recommend that you use Hive on Cassandra, mapping your Cassandra tables to Hive tables, usig the connector: cassandra_handler_for_hive, ( and then use hive to do joins and conditions on non partition keys)
I should mention that the performance of queries using Hive with Cassandra is not reasonable, (I have had a case where count(*) on a table with 500M records took 1 hour on 4 nodes). As a work around I used to copy the tables in HDFS after that do the computation using data on HDFS, but this is not good solution if you are seeking the fresh data.
Now for your question: To use Cassandra or not, it depends realy on your needs, Cassandra have a good performance in read/write per second record.
If your needs are met with using Hive/Cassandra to do the queries you need, so why not stay on Cassandra?
For a complex real-time Apache Storm topology I need aggregates of my data (stored in CassandraDB) for some computation steps. So far the data is queried when needed with CQL (Cassandra Query Language) and aggregated in a Storm bolt. That is a bit slow, so we want to have the data needed for the aggregation cached. Two option are on the table:
Put the data needed in an indexed Ignite Cache and sliding-window-query it from Storm. In this case we would only need one Cache and use different queries, depending on the aggregation.
Put the data in Cassandras in-memory, off-heap cache.
Argument for Ignite: We only need one indexed cache, while we would need one Cassandra table for each aggregation, for fast access. (Also ACID, but obviously we already live with CAP, so not a strong argument for our architects.)
Argument for Cassandra: We don't need to introduce a new technology.
But: What about speed? How fast would an indexed Ignite cache be compared to an optimized (= own table for each query) in-memory Cassandra?
I believe that in-memory indexed SQL in Ignite would be faster than Cassandra CQL queries. Apache Ignite is ANSI-99 SQL compatible, so you should be able to do all sorts of aggregations, joins, order by, group by, etc.
I will raise a point within the Ignite community to see if Cassandra CQL could be benchmarked against Ignite SQL. When done, will post the results here.