When trades_to_price is given start-argument of numpy.int64 type instead of the base int type, "buffer size mismatch" exception is raised. Int64 from numpy apparently appear very commonly in datetime to epoch conversions.
Example: using trades within a set number of hours prior to the latest timestamp;
When
trades_to_price
is given start-argument ofnumpy.int64
type instead of the baseint
type, "buffer size mismatch" exception is raised. Int64 from numpy apparently appear very commonly in datetime to epoch conversions.Example: using trades within a set number of hours prior to the latest timestamp;
This raises an error unless:
Suggestion: Include
int()
conversion in thesql_to_df
function ifstart
argument is not a string, around here.PS: Took some debugging since the error looked quite obscure (at least for a non-python person).