Skip to content

@filter & @sort

Filter rows with SQL-like conditions and sort DataFrames by one or more columns.


@filter

Simple Example

import additory as add
import polars as pl

df = pl.DataFrame({
    'name': ['Alice', 'Bob', 'Carol', 'Dave'],
    'age': [25, 17, 30, 15],
    'department': ['Sales', 'HR', 'Sales', 'HR'],
})

result = add.transform('@filter', df, where='age >= 18')
print(result)

Parameters

Parameter Type Default Description
where str None SQL-like filter condition
columns str or list[str] None Select specific columns in the output

Filter Conditions

The where parameter accepts SQL-like expressions:

# Comparison operators
result = add.transform('@filter', df, where='age >= 18')
result = add.transform('@filter', df, where='salary > 50000')
result = add.transform('@filter', df, where='status == "active"')

# Null checks
result = add.transform('@filter', df, where='email is not null')
result = add.transform('@filter', df, where='phone is null')

# String matching
result = add.transform('@filter', df, where='name == "Alice"')

Column Selection

Use columns to select specific columns in the output (independent of row filtering):

# Select columns only
result = add.transform('@filter', df, columns=['name', 'age'])

# Filter rows AND select columns
result = add.transform('@filter', df, where='age >= 18', columns=['name', 'department'])

@sort

Simple Example

import additory as add
import polars as pl

df = pl.DataFrame({
    'name': ['Carol', 'Alice', 'Bob'],
    'score': [85, 92, 78],
})

result = add.transform('@sort', df, by='score', strategy='desc')
print(result)

Parameters

Parameter Type Default Description
by str or list[str] (required) Column(s) to sort by
strategy str 'asc' Sort direction: 'asc' (ascending) or 'desc' (descending)

Ascending and Descending

# Ascending (default)
result = add.transform('@sort', df, by='name')

# Descending
result = add.transform('@sort', df, by='score', strategy='desc')

Multi-column Sort

Sort by multiple columns in priority order:

df = pl.DataFrame({
    'department': ['Sales', 'HR', 'Sales', 'HR'],
    'name': ['Carol', 'Alice', 'Bob', 'Dave'],
    'salary': [60000, 55000, 70000, 50000],
})

result = add.transform('@sort', df, by=['department', 'salary'], strategy='asc')

Practical Scenarios

Filter then sort

import additory as add
import polars as pl

employees = pl.DataFrame({
    'name': ['Alice', 'Bob', 'Carol', 'Dave', 'Eve'],
    'department': ['Engineering', 'Sales', 'Engineering', 'HR', 'Sales'],
    'salary': [90000, 60000, 85000, 55000, 70000],
})

# Filter to engineering, then sort by salary descending
engineers = add.transform('@filter', employees, where='department == "Engineering"')
ranked = add.transform('@sort', engineers, by='salary', strategy='desc')
print(ranked)

Column selection for reporting

# Get just names and departments, sorted
report = add.transform('@filter', employees, columns=['name', 'department'])
report = add.transform('@sort', report, by='name')

Next Steps