"In the csv file that value is represented as 1/0 and when it goes into
phoenix it should go as true/false. but it is going as false/true" > create table booltest(pk varchar not null primary key, truth boolean);
> upsert into booltest values('true', 1);
Error: ERROR 203 (22005): Type mismatch. INTEGER cannot be coerced to BOOLEAN (state=22005,code=203) It seems like UPSERTS will not coerce integers into a boolean. I'm curious how the CSV tool is doing this. psql.py seems to do this fine: $ echo "true,1" > ~/booleans.csv
$ echo "false,0" >> ~/booleans.csv
$ /usr/local/lib/phoenix/bin/psql.py -t BOOLTEST localhost:2181:/hbase-1.2 ~/booleans.csv
$ sqlline.py ...
> 0: jdbc:phoenix:localhost:2181:/hbase-1.2> select * from booltest;
+--------+--------+
| PK | TRUTH |
+--------+--------+
| false | false |
| true | true |
+--------+--------+
Similarly, using the CsvBulkLoadTool shows the same for me: yarn jar /usr/local/lib/phoenix/phoenix-4.9.0-HBase-1.2-SNAPSHOT-client.jar org.apache.phoenix.mapreduce.CsvBulkLoadTool -Dmapred.map.child.java.opts="-Xmx1G" --table BOOLTEST --input booleans.csv -z localhost:2181:/hbase-1.2 Maybe do a sanity check on your processing?
... View more
