SemanticException Partition spec {col=null} contai

2019-06-26 09:14发布

问题:

I am trying to create dynamic partitions in hive using following code.

SET hive.exec.dynamic.partition = true;
SET hive.exec.dynamic.partition.mode = nonstrict;

create external table if not exists report_ipsummary_hourwise(
ip_address string,imp_date string,imp_hour bigint,geo_country string)
PARTITIONED BY (imp_date_P string,imp_hour_P string,geo_coutry_P string) 
row format delimited 
fields terminated by '\t'
stored as textfile
location 's3://abc';

insert overwrite table report_ipsummary_hourwise PARTITION (imp_date_P,imp_hour_P,geo_country_P)
SELECT ip_address,imp_date,imp_hour,geo_country,
       imp_date as imp_date_P,
       imp_hour as imp_hour_P,
       geo_country as geo_country_P
FROM report_ipsummary_hourwise_Temp;

Where report_ipsummary_hourwise_Temp table contains following columns, ip_address,imp_date,imp_hour,geo_country.

I am getting this error

SemanticException Partition spec {imp_hour_p=null, imp_date_p=null, geo_country_p=null} contains non-partition columns.

Can anybody suggest why this error is coming ?

回答1:

You insert sql have the geo_country_P column but the target table column name is geo_coutry_P. miss a n in country



回答2:

I was facing the same error. It's because of the extra characters present in the file. Best solution is to remove all the blank characters and reinsert if you want.



回答3:

It says while copying the file from result to hdfs jobs could not recognize the partition location. What i can suspect you have table with partition (imp_date_P,imp_hour_P,geo_country_P) whereas job is trying to copy on imp_hour_p=null, imp_date_p=null, geo_country_p=null which doesn't match..try to check hdfs location...the other point what i can suggest not to duplicate column name and partition twice



回答4:

It could also be https://issues.apache.org/jira/browse/HIVE-14032

INSERT OVERWRITE command failed with case sensitive partition key names

There is a bug in Hive which makes partition column names case-sensitive.

For me fix was that both column name has to be lower-case in the table and PARTITION BY clause's in table definition has to be lower-case. (they can be both upper-case too; due to this Hive bug HIVE-14032 the case just has to match)