Expressions

FlowFrame methods accept standard Polars expressions — ff.col, operators, ff.when, and most of the Polars expression API. Expressions are the default way to express transformations.

Most of Polars, not all of it

Nearly every Polars Expr method is available, but a few names track the pinned Polars version (cum_sum, not cumsum; dt.weekday, not day_of_week), some helpers are selectors rather than top-level functions (ff.all_(), not ff.all(); use ff.col("*").exclude(...), there is no ff.exclude()), and expressions without a dedicated node render as polars_code nodes in the visual editor. Which operations become which node lists the patterns that render natively.

Column references and arithmetic

import flowfile as ff

df = df.with_columns([
    (ff.col("price") * ff.col("quantity")).alias("revenue"),
    (ff.col("price") * 1.1).alias("price_with_tax"),
    (ff.col("total") / ff.col("count")).alias("average"),
])

Conditional logic

df = df.with_columns(
    ff.when(ff.col("quantity") > 75)
    .then(ff.lit("High"))
    .otherwise(ff.lit("Low"))
    .alias("volume_category")
)

A when chain renders as a Formula node with the equivalent if … then … elseif … else … endif formula, so the branches stay editable in the visual editor. Membership tests use is_in: ff.when(ff.col("team").is_in(["DS", "DE"])) becomes if [team] in ("DS", "DE") then … endif. Conditions without a formula form (a map_elements call, for example) push the whole chain to a polars_code node.

Filtering

df = df.filter(ff.col("price") > 100)
df = df.filter(ff.col("status") != "cancelled", description="Drop cancelled orders")

Comparison predicates render as a Filter node whose advanced expression is the equivalent formula: the first call above becomes ([price] > 100), and ff.col("status").is_in(["new", "open"]) becomes [status] in ("new", "open"). Several predicates are joined with and. A predicate without a formula form (a lambda, or a method the Formula Language does not cover) renders as a polars_code node instead.

Namespaces

Polars expression namespaces work as expected:

df = df.with_columns([
    ff.col("name").str.to_uppercase().alias("name_upper"),
    ff.col("order_date").dt.year().alias("order_year"),
])

Because expressions stay lazy, they compose into one query plan that runs when you .collect() — see FlowFrame and FlowGraph.

Looking for the Excel-like [column] syntax?

That is the Flowfile formula language — a separate, simpler syntax shared with the visual editor. See Formulas in Python for the FlowFrame methods that accept it, and the Formula Language guide for the language itself.