[082349872349872]

> If you wanted to operate on or compare more than one set, you had to write your own functions.

This may be a 20th century viewpoint, but I would hesitate to call anyone who couldn't write all of these functions a programmer.

Introducing network round trips to use convenient syntax seems ... suboptimal.

[ringofchaos]

Using pandas or numpy is more than uisng optimal syntax

There is a reason why python is used for ML for mathematical calculations. You could do all low level matrix multiplication using custom functions uisng any language. But is that optimal way to dob that.

My use case is similar

[082349872349872]

Fair enough; where's the breakeven?