Module `Agg.Int`

Integer aggregations - work on any integer column type.

Operations in this module accept int32 and int64 columns only. Results preserve integer semantics where possible (sum returns int64, mean returns float since averages are often fractional).

raises Invalid_argument
if column is not an integer type.

val sum : t -> string -> int64

sum df name returns sum as int64.

Works on int32 and int64 columns. Uses int64 to avoid overflow issues that could occur with int32 sums. Null values (as indicated by the column mask) are excluded from the sum.

Time complexity: O(n) where n is the number of rows.

val min : t -> string -> int64 option

min df name returns minimum value as int64, None if empty or all nulls.

Sentinel null values (Int32.min_int, Int64.min_int) are excluded from comparison. Result is converted to int64 for consistency.

Time complexity: O(n) where n is the number of rows.

val max : t -> string -> int64 option

max df name returns maximum value as int64, None if empty or all nulls.

Sentinel null values are excluded from comparison.

Time complexity: O(n) where n is the number of rows.

val mean : t -> string -> float

mean df name returns mean as float.

Since the mean of integers is often fractional, the result is always a float regardless of input type. Null values are excluded from calculation.

Time complexity: O(n) where n is the number of rows.