Skip to content

Reinterpret the underlying bits of a same-size numeric type.

Source code

Description

This operation is only allowed between numeric types of the same size. To convert between types of different sizes, use <expr>$cast() instead.

Usage

<Expr>$reinterpret(..., signed = NULL, dtype = NULL)

Arguments

… These dots are for future extensions and must be empty.
signed Whether to reinterpret as a signed or unsigned integer of the same width as the input. Must specify exactly one of signed and dtype.
dtype Target data type for the reinterpretation. Must specify exactly one of signed and dtype.

Details

A Decimal can also be reinterpreted to and from Int128, its unscaled integer representation: 1.23 as Decimal(5, 2) becomes 123. Reinterpreting an Int128 raises an error if the value doesn’t fit the precision of the target Decimal.

Value

A polars expression

Examples

library("polars")

df <- pl$DataFrame(a = c(1, 1, 2))$cast(pl$UInt64)

# Reinterpret column a as Int64
df$with_columns(
  reinterpreted = pl$col("a")$reinterpret(dtype = pl$Int64)
)
#> shape: (3, 2)
#> ┌─────┬───────────────┐
#> │ a   ┆ reinterpreted │
#> │ --- ┆ ---           │
#> │ u64 ┆ i64           │
#> ╞═════╪═══════════════╡
#> │ 1   ┆ 1             │
#> │ 1   ┆ 1             │
#> │ 2   ┆ 2             │
#> └─────┴───────────────┘
# Reinterpret a Decimal as its unscaled Int128 representation
df2 <- pl$DataFrame(a = c(1.23, -4.5))$cast(pl$Decimal(5, 2))
df2$with_columns(
  reinterpreted = pl$col("a")$reinterpret(dtype = pl$Int128)
)
#> shape: (2, 2)
#> ┌──────────────┬───────────────┐
#> │ a            ┆ reinterpreted │
#> │ ---          ┆ ---           │
#> │ decimal[5,2] ┆ i128          │
#> ╞══════════════╪═══════════════╡
#> │ 1.23         ┆ 123           │
#> │ -4.50        ┆ -450          │
#> └──────────────┴───────────────┘