tddbapi

tddbapi is a DB API module that is a shim over teradatasql, the Teradata provided DB API module. teradatasql closely follows Teradata JDBC driver in terms of functionality and behaviors, some of which, arguably aren't quite pythonic. tddbapi attempts to correct this by altering selected teradatasql behavior (APIs) as documented here. The remaining, unaltered, functionality is directly passed through to the teradatasql module.

teradatasql behavior

1. Fixed-length character column values have incorrect length when accessed in Python. According to Teradata, this is working as designed.

2. teradatasql uses OperationalError for all exceptions. According to Python's DB API specification, OperationalError is reserved for a certain class of errors. Additionally, the exception object

   • doesn't provide easy access to the error code or the error message values.
   • has text that includes unformatted stack trace of underlying gosql driver. While this may be useful in certain situations, for example, reporting errors back to Teradata, but may not be the best message to display to the users.

3. All SQL INTERVAL values are returned as Python type str instead of more appropriate native Python types, thus working with them is harder in Python. Apparently, this was done because not all SQL INTERVAL data types can be represented by Python's datetime.timedelta type.

4. teradatasql exposes all internally used names, and thus can pollute the namespace if you use from teradatasql import *.

tddbapi behavior

1. Fixed-length character columns will have any extra spaces beyond their actual definition truncated.

2. Appropriate Exception raised for the most common errors. Exception objects:

   • have sqlcode and sqltext attributes that can be queried directly instead of having to parse the error message
   • provide __str__ method that produces a human-readable error message and removes gosql stack-trace information
   • note: If you want to ensure that the exception handling logic works with teradatasql and tddbapi, use an Exception class that is higher in the hierarchy, for example, DatabaseError or Error

3. INTERVAL values are objects of either datetime.timedelta subclasses, or for the month based interval types, instances of int subclasses. The month-based intervals use the total number of months as canonical value, thus allowing comparison between two month-based intervals even if their exact types are different.

4. It's safe to code from tddbapi import * without introducing non DB-API names into the current namespace

Note tddbapi always obtains resultset metadata information for all queries. If your application also needs access to resultset metadata, it can be accessed as _meta attribute of the cursor as already-parsed JSON data.

Installation and Usage

Install tddbapi using the standard pip utility:

pip install --upgrade tddbapi

Using tddbapi in a Python application is as simple as replacing teradatasql with tddbapi. For example:

import tddbapi as teradatasql

with teradatasql.connect(host="whomooz", user="guest", password="please") as con:
    with con.cursor() as cur:
        cur.execute("SELECT * FROM DBC.DBCINFOV")
        print(cur.fetchall())

Disclaimers

tddbapi module:

1. does not come with any warranty and the user assumes all risks associated with its usage.
2. is not endorsed by Teradata.
3. overrides some internal teradatasql implementation details. It is possible that a future version of teradatasql can break tddbapi, at least for a short period.
4. will be kept compatible with teradatasql on a best-effort basis. If you must have the latest teradatasql all the time and cannot live with the risk of breakage, even for a short period, you shouldn't use this module.
5. is available via MIT license, which may be different from the licenses of any of its dependencies (currently only teradatasql).