Set argument for redshift connection/cursor in order to fetch float instances?

[ni-todo-spot]

Driver version

2.0.905

Redshift version

PostgreSQL 8.0.2 on i686-pc-linux-gnu, compiled by GCC gcc (GCC) 3.4.2 20041017 (Red Hat 3.4.2-6.fc3), Redshift 1.0.36224��

Client Operating System

Darwin Nativs-MacBook-Pro.local 21.4.0 Darwin Kernel Version 21.4.0: Mon Feb 21 20:34:37 PST 2022; root:xnu-8020.101.4~2/RELEASE_X86_64 x86_64

Python version

3.8.7

Problem description

I was wondering if there is some kind of a flag / setting I can set in order to tell the connector return a numeric result as float / np.float instead of decimal.Decimal instance?

I was using the following raw sql on a cursor instance:

cursor.execute("SELECT 1.0 / 7 AS foo")
df = cursor.fetch_dataframe()
output = df.squeeze()
assert not isinstance(output, decimal.Decimal)

Thanks in advance!

[Brooke-white]

Hey @ni-todo-spot ,

Thanks for reaching out, this is a good question. At this time, all NUMERIC values are mapped to decimal.Decimal. decimal.Decimal is used for NUMERIC values as it offer a higher level of precision than float.

I will chat with the Redshift driver team to see if our Redshift JDBC driver offers such an option, and see if this is something we could support. Thank you for your patience :)

[Brooke-white]

Hi @ni-todo-spot ,

Small update -- I've scoped work required for this feature and it looks minimal. I'm working on implementation/testing and anticipate we can include this feature in our next release (scheduled for mid-April). At this time, I think supporting an option for returning a numeric as float would be most straight forward.

[ni-todo-spot]

Hi @ni-todo-spot ,

Small update -- I've scoped work required for this feature and it looks minimal. I'm working on implementation/testing and anticipate we can include this feature in our next release (scheduled for mid-April). At this time, I think supporting an option for returning a numeric as float would be most straight forward.

Great news! Thank you very much :)

[Brooke-white]

Hi @ni-todo-spot ,

We've just released v2.0.906 of redshift_connector which includes this feature. Please be sure to read the disclaimer about this flag, numeric_to_float located in the project README before using. I've copied it below for your convenience. This disclaimer is just warning of the precision differences between decimal.Decimal and float, to ensure user is aware of this before use :).

Specifies if NUMERIC datatype values will be converted from decimal.Decimal to float. By default NUMERIC values are received as decimal.Decimal. Enabling this option is not recommended for use cases which prefer the most precision as results may be rounded. Please reference the Python docs on decimal.Decimal to see the tradeoffs between decimal.Decimal and float before enabling this option.

Please let me know if you have any questions by opening a new issue. Have a nice weekend! :)

[ni-todo-spot]

@Brooke-white Wow! 👏
This is what I've been waiting for :)

Thank you very much 😄