Operators

[Polars Documentation]

All Polars conjunction, comparison, and binary operators in the linked documentation are supported and are considered row-by-row.

Even if you are in an aggregation context like .select or .agg, OpenDP enforces that inputs to binary operators are row-by-row. This is to ensure that the left and right arguments of binary operators have meaningful row alignment.

These operators are particularly useful for building filtering predicates and grouping columns.

import polars as pl
import opendp.prelude as dp
dp.enable_features("contrib")
# Fetch and unpack the data.
![ -e ../sample_FR_LFS.csv ] || ( curl 'https://github.com/opendp/dp-test-datasets/blob/main/data/sample_FR_LFS.csv.zip?raw=true' --location --output sample_FR_LFS.csv.zip; unzip sample_FR_LFS.csv.zip -d ../ )

context = dp.Context.compositor(
    # Many columns contain mixtures of strings and numbers and cannot be parsed as floats,
    # so we'll set `ignore_errors` to true to avoid conversion errors.
    data=pl.scan_csv("../sample_FR_LFS.csv", ignore_errors=True),
    privacy_unit=dp.unit_of(contributions=36),
    privacy_loss=dp.loss_of(epsilon=1.0, delta=1e-7),
    split_evenly_over=1,
    margins={(): dp.polars.Margin(max_partition_length=60_000_000 * 36)}
)

query = (
    context.query()
    .filter((pl.col.HWUSUAL > 0) & (pl.col.HWUSUAL != 99))  # using the .gt, .and_ and .ne operators
    .with_columns(OVER_40=pl.col.AGE > 40)
    .group_by("SEX", "OVER_40")
    .agg(dp.len())
)
query.release().collect().sort("SEX", "OVER_40")

shape: (4, 3)

SEX	OVER_40	len
i64	bool	u32
1	false	18045
1	true	22883
2	false	15838
2	true	21500