Problem creating external table to a sequence file on HDFS

Hi I am trying to create a external table to a sequence file on HDFS. I have my own input format and a SerDe, which is compiled into a jar and added in HIVE. Inspite of doing this, I get the following error. Please Suggest. hive create table seq (key STRING, value STRING) ROW FORMAT SERDE

Where should Hive Process Installed??

Hi I am using Hive-0.9.0 + Hadoop-2.0.1 with 2 machine. One machine contains say Machine-1 : NameNode, SecondaraNameNode , ResourceManager and Hive Machine-2 : Proxy server, JHS , DataNode and NodeManager. Problem : When I execute Job queries i.e “select count(key) from src” for the

Re: Hive Problems Reading Avro+Snappy Data

Thanks Chuck. I think the problem is the job configuration on the query. I logged back into the system this morning and started a new Hive client shell and issued a series of more complex queries against the Avro+Snappy table and they all worked fine. So I started trying to recall what could

RE: Hive Problems Reading Avro+Snappy Data

Just so you know there is still at least one bug using avro+compression like snappy: https://issues.apache.org/jira/browse/HIVE-3308 There's a simple one line patch but unfortunately it's not committed yet. From: Thomas, Matthew [mailto:mtho...@verisign.com] Sent: Monday, April 08, 2013 1:59 PM

StackOverflowError when add jar using multiple thread cocurrently

Hi All, Recently we find that when multiple jdbc connection concurrently add jars, hiveserver will throw StackOverflowError when serializeMapRedWork to hdfs, I find the relate issue hive-2666 is similar, but I think it missed the concurrently scenario. I find it is because the classloader is

RE: Syntax for filters on timstamp data type

Hi Mark, Correct, I just did some tests and the cast is the way to go. While for comparison operations (equal, diff, ...) implicit casts work, this is not the case for the IN clause. I think it should, as eventually this just translates to a disjunction of comparisons so it should be the same.

Re: Syntax for filters on timstamp data type

great so can we assume that equals comparison can auto cast but not the in range statement ? On Mon, Apr 8, 2013 at 7:08 PM, LUTTER, Steffen steffen.lut...@sap.comwrote: Hi Mark, ** ** Correct, I just did some tests and the cast is the way to go. While for comparison operations

RE: Syntax for filters on timstamp data type

Exactly From: Nitin Pawar [mailto:nitinpawar...@gmail.com] Sent: 08 April 2013 16:06 To: user@hive.apache.org Subject: Re: Syntax for filters on timstamp data type great so can we assume that equals comparison can auto cast but not the in range statement ? On Mon, Apr 8, 2013 at 7:08 PM,

Re: Huge join performance issue

Did you verify that all your available mappers are running (and reducers too)? If you have a small number of partitions with huge files, you might me underutilizing mappers (check that the files are being split). Also, it might be optimal to have a single wave of reducers by setting the number of

Should the SQL schema for HIVE be identical to the SQL schema for a relational database?

Hi It would be terrific to get some advice on migrating a schema from RDMS to Hive. Should the SQL schema for HIVE be identical to the SQL schema for a Posgresql/mysql database? Specifically: I have an application that generates events that look like the following:

RE: Should the SQL schema for HIVE be identical to the SQL schema for a relational database?

Some general guidance would be to aim for minimal JOINs in your regular queries. Thus the biggest change from a normal-form RDBMS schema is to denormalize such that joins do not come in to play until a query has already performed data reduction via filtering or aggregation. This implies a star