Reply
Explorer
Posts: 6
Registered: ‎11-16-2016

Unable to load data into Impala table stored as Kudu

I've created a table in Impala, stored as Kudu using Hue interface and trying to load data into the same from a csv file located on my HDFS folder.

 

However, the same is failing with the error message:

AnalysisException: LOAD DATA only supported for HDFS tables: bit_nnet.feature_matrix_100

 

The following is the CREATE TABLE command using:

create table if not exists bit_nnet.feature_matrix_100 (
	tan							string		not null,
	...
	primary key(tan)
)
partition by hash(tan) partitions 4
stored as kudu;

 

The following is the LOAD DATA command I am using:

load data inpath '/data/feature_matrix_100.csv' into table bit_nnet.feature_matrix_100;
Cloudera Employee
Posts: 47
Registered: ‎02-05-2016

Re: Unable to load data into Impala table stored as Kudu

As you saw in the error message, the "LOAD DATA" statement doesn't work for Kudu tables. Its documentation says as much.

 

I'm no Impala expert, but perhaps you can build an HDFS table around '/data/feature_matrix_100.csv' first, then SELECT that data into your Kudu table?

New Contributor
Posts: 1
Registered: ‎12-02-2018

Re: Unable to load data into Impala table stored as Kudu

Did you have solve this problem? I encounterd this also. Could you share your experience?

Highlighted
Champion
Posts: 753
Registered: ‎05-16-2016

Re: Unable to load data into Impala table stored as Kudu

@xuehan2216 @sdutta007 

 

You can perform CTAS - please refer the below reference 

CREATE TABLE new_table
PRIMARY KEY (ts, name)
PARTITION BY HASH(name) PARTITIONS 8
STORED AS KUDU
AS SELECT ts, name, value FROM old_table;
Announcements