georgia-tech-db / eva / #913

0%

master: 0%

push

circleci

Remove table names from column names for `df()` call (#1256)

Removing table names from the `dataframe` during `df()` call. The users
can then easily load CSV files generated using `EvaDB` with the
`to_csv()` call at a later time (for long-running or expensive queries).

Example:

```
select_query = cursor.query(
    f"SELECT * FROM {repo_name}_StargazerList;"
).df()

select_query.to_csv("stargazers_list.csv", index=False)

# Later
cursor.query(
        f"""
   CREATE TABLE IF NOT EXISTS {repo_name}_StargazerList(
   github_username TEXT(1000));
"""
    ).df()

cursor.query("LOAD CSV 'stargazers_list.csv' INTO {repo_name}_StargazerList;""").df()

```

Do we need the table names for any use cases? For example, for duplicate
column names from two different functions - `object_detector_1.labels`
and `object_detector_2.labels`?

---------

Co-authored-by: Andy Xu <xzdandy@example.com>
Co-authored-by: Andy Xu <xzdandy@gmail.com>

2 of 7 new or added lines in 1 file covered. (28.57%)

9125 of 12498 relevant lines covered (73.01%)

0.73 hits per line