Alternative method 10x faster than dt.offset_by()

[Chuck321123]

Alternative method 10x faster than dt.offset_by()

[MarcoGorelli]

thanks @Chuck321123 for the report

agree - I think something similar to #16615 and #16666 could be done

will try to work on a pr this week

[Chuck321123]

@MarcoGorelli Thanks. Based of this example
%timeit df.with_columns(pl.from_epoch((pl.col("UTC_Time").dt.epoch(time_unit="us") + 5 * 60 * 1_000_000), time_unit="us"))
notice how much slower it suddenly became when I switched to "us" (or anything other than "ns")
3.87 ms ± 142 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) If we see issue case #16716, we see similar differences where operations done on anything other than nanoseconds, performs much slower. Don't know if this is expected, but if it's not, it might be worth looking into this as well