Keywords in column names are not supported due to missing escaping support

[Christoph-W]

Given a database system view with column names being keywords can not be handled by JDBDT.
It is possible to create table tables meta data:

partitionsTable = JDBDT.table("USER_TAB_PARTITIONS")
    .columns("TABLE_NAME", "PARTITION_NAME", "HIGH_VALUE", "\"INTERVAL\"")
    .build(database)

But using it during data preparation

JDBDT.builder(partitionsTable)
    .sequence(...)
    .nullValue("\"INTERVAL\"")

fails with org.jdbdt.InvalidOperationException: Invalid column name: '"INTERVAL"'.
Using .remainingColumnsNull() does not reuse the metadata correctly ending in not escaped column name usage with Syntax error in SQL statement.
Also alternative escaping with ` fails. Even enabling of Option.CASE_SENSITIVE_COLUMN_NAMES did not help.

[edrdo]

@Christoph-W This is a tricky edge case where the behavior should be highly dependent on the database engine at stake; which one are you using? Have you tried using nullValue("INTERVAL") (I guess it should not work) ? Try also: 1) defining the column name in the definition of partitionsTable as "\"interval\"" rather than "\"INTERVAL\"" with Option.CASE_SENSITIVE_COLUMN_NAMES enabled and 2) later use "interval" in data set builder operations.

[Christoph-W]

The following combination seems to be the single valid solution:

database.enable(Option.CASE_SENSITIVE_COLUMN_NAMES);
partitionsTable = JDBDT.table("USER_TAB_PARTITIONS")
    .columns("TABLE_NAME", "PARTITION_NAME", "HIGH_VALUE", "\"interval\"")
    .build(database);
...
JDBDT.builder(partitionsTable)
    .sequence(...)
    .nullValue("\"interval\"")

It does not seem reasonable to me to discard the support of upper case keywords in column and/or table names.
Finally, I would be happy if I could find hints in the documentation that keywords can only be escaped if they are written in lower case.

[edrdo]

It does not seem reasonable to me to discard the support of upper case keywords in column and/or table names.
I would be happy if I could find hints in the documentation that keywords can only be escaped if they are written in lower case.

@Christoph-W Yes that's how it works and the documentation is not explicit. The overall scheme for escaped column names is currently fragile.

[edrdo]

@Christoph-W I have been examining this issue, trying to improve support for the upcoming 1.4.0 release but a small incremental fix at this time is not possible. The overall scheme will have to be re-examined in a future release.