Bill Oliver created HIVE-6969:
---------------------------------
Summary: SQL literal should be a fixed length, not string
Key: HIVE-6969
URL: https://issues.apache.org/jira/browse/HIVE-6969
Project: Hive
Issue Type: Bug
Affects Versions: 0.12.0
Environment: I tried this with Hive .12 using cdh5.0.
Reporter: Bill Oliver
I checked to see if this was already reported, but could not find it.
Consider this simple query with a SQL literal:
hive> explain select 'abc' from pplsqlb0;
OK
ABSTRACT SYNTAX TREE:
(TOK_QUERY (TOK_FROM (TOK_TABREF (TOK_TABNAME pplsqlb0))) (TOK_INSERT
(TOK_DESTINATION (TOK_DIR TOK_TMP_FILE)) (TOK_SELECT (TOK_SELEXPR 'abc'))))
STAGE DEPENDENCIES:
Stage-1 is a root stage
Stage-0 is a root stage
STAGE PLANS:
Stage: Stage-1
Map Reduce
Alias -> Map Operator Tree:
pplsqlb0
TableScan
alias: pplsqlb0
Select Operator
expressions:
expr: 'abc'
type: string
outputColumnNames: _col0
File Output Operator
compressed: false
GlobalTableId: 0
table:
input format: org.apache.hadoop.mapred.TextInputFormat
output format:
org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat
serde: org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe
Stage: Stage-0
Fetch Operator
limit: -1
'abc' is treated as a STRING type. It would be better if it were handled as a
CHAR(3) type. If the length > 255, I would like to see it as a VARCHAR(n) type.
Also, string functions based on the literal, such as TRIM('abc') should reflect
the correct length. SUBSTRING('abc', 1, 1) should return CHAR(1).
Thanks for any help on this!
-bill
--
This message was sent by Atlassian JIRA
(v6.2#6252)
