使用Pyhive插入Hive会引发错误

Question

I am using pyhive to interact with hive.

The SELECT statement going well using this code bellow.

# Import hive module and connect
from pyhive import hive
conn = hive.Connection(host="HOST")
cur = conn.cursor()

# Import pandas
import pandas as pd

# Store select query in dataframe 
all_tables = pd.read_sql("SELECT * FROM table LIMIT 5", conn)
print all_tables

# Using curssor 
cur = conn.cursor()
cur.execute('SELECT * FROM table LIMIT 5')
print cursor.fetchall()

Until here there is no problem. When I want to INSERT into hive.

Let's say I want to excute this query : INSERT INTO table2 SELECT Col1, Col2 FROM table1;

I tried :

cur.execute('INSERT INTO table2 SELECT Col1, Col2 FROM table1')

I recieve this error

pyhive.exc.OperationalError: TExecuteStatementResp(status=TStatus(errorCode=1, errorMessage=u'Error while processing statement: FAILED: Execution Error, return code 1 from org.apache.hadoop.hive.ql.exec.tez.TezTask', sqlState=u'08S01', infoMessages=[u'*org.apache.hive.service.cli.HiveSQLException:Error while processing statement: FAILED: Execution Error, return code 1 from org.apache.hadoop.hive.ql.exec.tez.TezTask:28:27', u'org.apache.hive.service.cli.operation.Operation:toSQLException:Operation.java:388', u'org.apache.hive.service.cli.operation.SQLOperation:runQuery:SQLOperation.java:244', u'org.apache.hive.service.cli.operation.SQLOperation:runInternal:SQLOperation.java:279', u'org.apache.hive.service.cli.operation.Operation:run:Operation.java:324', u'org.apache.hive.service.cli.session.HiveSessionImpl:executeStatementInternal:HiveSessionImpl.java:499', u'org.apache.hive.service.cli.session.HiveSessionImpl:executeStatement:HiveSessionImpl.java:475', u'sun.reflect.GeneratedMethodAccessor81:invoke::-1', u'sun.reflect.DelegatingMethodAccessorImpl:invoke:DelegatingMethodAccessorImpl.java:43', u'java.lang.reflect.Method:invoke:Method.java:498', u'org.apache.hive.service.cli.session.HiveSessionProxy:invoke:HiveSessionProxy.java:78', u'org.apache.hive.service.cli.session.HiveSessionProxy:access$000:HiveSessionProxy.java:36', u'org.apache.hive.service.cli.session.HiveSessionProxy$1:run:HiveSessionProxy.java:63', u'java.security.AccessController:doPrivileged:AccessController.java:-2', u'javax.security.auth.Subject:doAs:Subject.java:422', u'org.apache.hadoop.security.UserGroupInformation:doAs:UserGroupInformation.java:1698', u'org.apache.hive.service.cli.session.HiveSessionProxy:invoke:HiveSessionProxy.java:59', u'com.sun.proxy.$Proxy33:executeStatement::-1', u'org.apache.hive.service.cli.CLIService:executeStatement:CLIService.java:270', u'org.apache.hive.service.cli.thrift.ThriftCLIService:ExecuteStatement:ThriftCLIService.java:507', u'org.apache.hive.service.rpc.thrift.TCLIService$Processor$ExecuteStatement:getResult:TCLIService.java:1437', u'org.apache.hive.service.rpc.thrift.TCLIService$Processor$ExecuteStatement:getResult:TCLIService.java:1422', u'org.apache.thrift.ProcessFunction:process:ProcessFunction.java:39', u'org.apache.thrift.TBaseProcessor:process:TBaseProcessor.java:39', u'org.apache.hive.service.auth.TSetIpAddressProcessor:process:TSetIpAddressProcessor.java:56', u'org.apache.thrift.server.TThreadPoolServer$WorkerProcess:run:TThreadPoolServer.java:286', u'java.util.concurrent.ThreadPoolExecutor:runWorker:ThreadPoolExecutor.java:1149', u'java.util.concurrent.ThreadPoolExecutor$Worker:run:ThreadPoolExecutor.java:624', u'java.lang.Thread:run:Thread.java:748'], statusCode=3), operationHandle=None)

If I excute the same query in hive directly everything run well. Any thoughts?

NB: All my tables are external

CREATE EXTERNAL TABLE IF NOT EXISTS table ( col1 String, col2 String) stored as orc LOCATION 's3://somewhere' tblproperties ("orc.compress"="SNAPPY");

Answer 1

The solution was to add the username in the connection line; conn = hive.Connection(host="HOST", username="USER")

From what I understand hive queries divided on many type of operations (jobs). While you are performing a simple query ( ie. SELECT * FROM table ) This reads data from the hive metastore no mapReduce job or tmp tables needed to perform the query. But as soon as you switch to more complicated queries (ie. using JOINs) you end up having the same error.

The file code looks like this:

# Import hive module and connect
from pyhive import hive
conn = hive.Connection(host="HOST", username="USER")
cur = conn.cursor()
query = "INSERT INTO table2 SELECT Col1, Col2 FROM table1"
cur.execute(query)

So maybe it needs permission or something.. I will search more about this behavior and update the answer.

Answer 2

I'm not sure how to insert a pandas df using pyhive, but if you have pyspark installed, one option is that you could convert to a spark df and use pyspark to do it.

from pyspark.sql import sqlContext
spark_df = sqlContext.createDataFrame(pandas_df)
spark_df.write.mode('append').saveAsTable('database_name.table_name')

Answer 3

You can do the following using spark .

from pyspark.sql import sqlContext

# convert the pandas data frame to spark data frame
spark_df = sqlContext.createDataFrame(pandas_df)

# register the spark data frame as temp table
spark_df.registerTempTable("my_temp_table")

# execute insert statement using spark sql
sqlContext,sql("insert into hive_table select * from my_temp_table")

This will insert data in your data frame to a hive table.

Hope this helps you