Description
JDBC thin driver will return wrong metadata for column type when user creates a table with Date and Decimal type.
Steps to reproduce.
1. Start one node and create table using this command
create table a(a varchar, b decimal,c date, primary key(a));
2. Run "!desc a" to show the metadata of table a
This results is as follows:
TABLE_CAT
TABLE_SCHEM PUBLIC
TABLE_NAME A
COLUMN_NAME A
DATA_TYPE 12
TYPE_NAME VARCHAR
COLUMN_SIZE null
BUFFER_LENGTH null
DECIMAL_DIGITS null
NUM_PREC_RADIX 10
NULLABLE 1
REMARKS
COLUMN_DEF
SQL_DATA_TYPE 12
SQL_DATETIME_SUB null
CHAR_OCTET_LENGTH 2147483647
ORDINAL_POSITION 1
IS_NULLABLE YES
SCOPE_CATLOG
SCOPE_SCHEMA
SCOPE_TABLE
SOURCE_DATA_TYPE null
IS_AUTOINCREMENT NO
IS_GENERATEDCOLUMN NO
TABLE_CAT
TABLE_SCHEM PUBLIC
TABLE_NAME A
COLUMN_NAME B
DATA_TYPE 1111
TYPE_NAME OTHER
COLUMN_SIZE null
BUFFER_LENGTH null
DECIMAL_DIGITS null
NUM_PREC_RADIX 10
NULLABLE 1
REMARKS
COLUMN_DEF
SQL_DATA_TYPE 1111
SQL_DATETIME_SUB null
CHAR_OCTET_LENGTH 2147483647
ORDINAL_POSITION 2
IS_NULLABLE YES
SCOPE_CATLOG
SCOPE_SCHEMA
SCOPE_TABLE
SOURCE_DATA_TYPE null
IS_AUTOINCREMENT NO
IS_GENERATEDCOLUMN NO
TABLE_CAT
TABLE_SCHEM PUBLIC
TABLE_NAME A
COLUMN_NAME C
DATA_TYPE 1111
TYPE_NAME OTHER
COLUMN_SIZE null
BUFFER_LENGTH null
DECIMAL_DIGITS null
NUM_PREC_RADIX 10
NULLABLE 1
REMARKS
COLUMN_DEF
SQL_DATA_TYPE 1111
SQL_DATETIME_SUB null
CHAR_OCTET_LENGTH 2147483647
ORDINAL_POSITION 3
IS_NULLABLE YES
SCOPE_CATLOG
SCOPE_SCHEMA
SCOPE_TABLE
SOURCE_DATA_TYPE null
IS_AUTOINCREMENT NO
IS_GENERATEDCOLUMN NO
Column b and c has the wrong DATA_TYPE and TYPE_NAME.
Attachments
Issue Links
- blocks
-
IGNITE-10314 Spark dataframe will get wrong schema if user executes add/drop column DDL
-
- Resolved
-
- links to
