SQL Refactor

This library contains a class, Refactor, for refactoring a SQL with a given knowledge of tables and column names mapping.

Structur of the knowledge

The knowledge is a dictionary containing the mapping of old_table-new_table and corresponding old_column-new_column.

The structure of the knowledge:

{
    "old_table_name":
    {
        "new_table": "new_table_name",
        "column_knowledge":
        {
            "column_1" : "new_column_1",
            "column_2" : "new_column_2",

        }
        "preserved" : boolean
    },
    .
    .
    .
}

How to use

Code sample:

from sql_refactor import Refactor

# refer to Structure of Knowledge
knowledge = KNOWLEDGE

refactor = Refactor(knowledge)
refactor.refactor('''SELECT column_1, column_2 FROM old_table_name''')

new_query = refactor.result()

Result:

SELECT
  new_column_1 AS column_1,
  new_column_2 AS column_2
FROM
  `new_table_name`

Limitation

This repo is under development which contains some limitations.

Limitations of the sql_parser:

1. MERGE statement is not supported yet

2. CREATE [TEMPORARY] FUNCTION is not supported yet

3. Function call without () is not supported.

   For instance: TIMESTAMP '2021-01-01' can not be parsed, expected TIMESTAMP('2021-01-01')

4. DECLARE statement is not supported yet

Limitations of the sql_refactor:

1. Has been tested on non-scripting SQL, and it works fine.
2. For the scripting SQL, only works if the query only contains CREATE TEMP TABLE + main select