AnsweredAssumed Answered

Drill on MapR-DB Binary Tables Failing Where/Like Clause?

Question asked by john.humphreys on Jun 13, 2017
Latest reply on Jun 19, 2017 by MichaelSegel

I created a MapR-DB binary table and populated it with around 7 million records in Spark.

 

The row key is of the format host#_metric#_timestamp.

 

When I try to select all data for host 1, metric 1, using a like clause, it returns far more records than it should (though it does return the correct ones too).  For example:

 

select CONVERT_FROM(row_key, 'UTF8')
from fullsparktest
where CONVERT_FROM(row_key, 'UTF8') like '1_1_%'

Returns 715,020 records.  Here is an excerpt showing some correct ones and some wrong ones.

1_19_1497382756
1_19_1497382816
1_1_1497381733
1_1_1497381793

am I doing something wrong here?

Outcomes