Finally Worked for Me and did some work around. Steps as below. 1) Create Temp table with same columns. 2) Overwrite table with required row data. 3)Drop Hive partitions and HDFS directory. 4)Insert records for respective partitions and rows. 5) verify the counts. 1) hive> select count(*) from emptable where od='17_06_30' and ccodee=!'123';
OK
27
hive> select count(*) from emptable where od='17_06_30' and ccodee='123';
OK
7
hive>show create table emptable_tmp; :- Note hdfs location
2)Create table and overwrite with required partitioned data
hive> CREATE TABLE `emptable_tmp`(
'rowid` string,PARTITIONED BY (`od` string)
ROW FORMAT SERDE
'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe'
STORED AS INPUTFORMAT
'org.apache.hadoop.mapred.SequenceFileInputFormat';
hive> insert into emptable_tmp partition(od) select * from emptable where od='17_06_30' and ccodee!='123';
Time taken for adding to write entity : 1
Partition database.emptable_tmp{od=17_06_30} stats: [numFiles=20, numRows=27,totalSize=6216,rawDataSize=5502]
OK
3) Drop Partitions from Hive and HDFS directory as well, as this is External table.
hive> alter table emptable drop partition(od='17_06_30');
Dropped the partition od=17_06_30
OK
Time taken: 0.291 seconds
HDFS partition deletion
#hdfs dfs -rm -r /hdfs/location/emptable/ods='17_06_30'
4) Insert data for that partition only.
hive> insert into emptable partition(od) select * from emptable_tmp;
Partition database.emptable{ds=17_06_30} stats: [numFiles=66, numRows=20, totalSize=5441469982, rawDataSize=]
OK
Time taken: 27.282 seconds
5) Verifying the counts on partitions and respective rows data
1) hive> select count(*) from emptable where od='17_06_30' and ccodee=!'123';
OK
27
hive> select count(*) from emptable where od='17_06_30' and ccodee='123';
OK
0
... View more
