Filter

Filter keeps rows that match your conditions and removes all others. Each filter targets one column, applies a condition (equal, contains, is empty, etc.), and specifies the value to compare against, if the condition needs one. Add multiple filters to narrow further. Every filter must be satisfied for a row to survive (AND logic).

All column types are supported: numeric, categorical (string), and datetime.

When to Use Filter

• Isolating a segment. Keep only orders from a specific region, product category, or customer tier for segment-level analysis.
• Removing test or invalid data. Exclude rows where status = "test" or revenue <= 0 before running aggregations.
• Date range scoping. Keep only rows within a specific date range before exporting or joining to another dataset.
• Finding missing data. Isolate rows where email is empty to route them to a cleanup step, or keep only rows where a required field is filled in.
• Partial text matching. Keep rows where customer_email contains @gmail.com, or where sku starts with a specific product-line prefix.
• Excluding known bad values. Remove rows where a category column contains a deprecated label you are not yet ready to replace with Find & Replace.
• Narrowing before Group By. Filter to the relevant subset before aggregating to avoid computing over the full dataset.

Sample Dataset

The examples in this doc use the Edilitics sample orders dataset. Download it to follow along in your own workspace.

Relevant columns for Filter examples:

Conditions by Column Type

The available conditions change depending on the type of column selected.

Multiple Filters and AND Logic

When you add more than one filter row, all filters are applied sequentially. A row must pass every filter to appear in the output. This is AND logic.

Example: two filters applied together:

• revenue Greater Than Equal to 1000
• region Equal to East

Only rows where revenue is 1000 or above AND region is East are kept. Rows matching only one condition are removed.

To implement OR logic (keep rows matching either condition), use the Code Editor operation with a Polars expression: df.filter((pl.col("region") == "East") | (pl.col("region") == "West")).

How to Apply Filter

Before and After

Two filters applied in one operation:

• revenue Greater Than Equal to 1000
• status Not Equal to cancelled

Input:

Output (rows passing both conditions):

3 rows removed: two for revenue below 1000, one for cancelled status.

Code Equivalent

-- AND logic: both conditions must be true
SELECT *
FROM orders
WHERE revenue >= 1000
  AND status != 'cancelled';

-- Date range filter
SELECT *
FROM orders
WHERE order_date >= '2024-01-01'
  AND order_date <= '2024-06-30';

-- Text match and missing values
SELECT *
FROM orders
WHERE customer_email LIKE '%@gmail.com%'
  AND email IS NOT NULL;

import polars as pl

# AND logic: chain .filter() calls
df = (
    df
    .filter(pl.col("revenue").ge(1000))
    .filter(~pl.col("status").is_in(["cancelled"]))
)

# Equal to multiple values
df = df.filter(pl.col("region").is_in(["East", "West"]))

# Date range
df = df.filter(
    pl.col("order_date").ge(pl.lit("2024-01-01").cast(pl.Date))
).filter(
    pl.col("order_date").le(pl.lit("2024-06-30").cast(pl.Date))
)

# Contains / starts with / ends with
df = df.filter(pl.col("customer_email").str.contains("@gmail.com", literal=True))
df = df.filter(pl.col("sku").str.starts_with("WH-"))
df = df.filter(pl.col("sku").str.ends_with("-CLR"))

# Is Empty / Is Not Empty
df = df.filter(pl.col("email").is_null())
df = df.filter(pl.col("email").is_not_null())

Operation Reference

Frequently Asked Questions

Next Steps

Need help? Email support@edilitics.com with your workspace, job ID, and context. We reply within one business day.