feat(connect): add df.filter

src/daft-connect/src/translation/logical_plan.rs

@@ -5,11 +5,12 @@
 use tracing::warn;
 
 use crate::translation::logical_plan::{
-    aggregate::aggregate, local_relation::local_relation, project::project, range::range,
-    read::read, to_df::to_df, with_columns::with_columns,
+    aggregate::aggregate, filter::filter, local_relation::local_relation, project::project,
+    range::range, read::read, to_df::to_df, with_columns::with_columns,
 };
 
 mod aggregate;
+mod filter;
 mod local_relation;
 mod project;
 mod range;
@@ -59,6 +60,9 @@
         RelType::Project(p) => project(*p)
             .await
             .wrap_err("Failed to apply project to logical plan"),
+        RelType::Filter(f) => filter(*f)
+            .await
+            .wrap_err("Failed to apply filter to logical plan"),
         RelType::Aggregate(a) => aggregate(*a)
             .await
             .wrap_err("Failed to apply aggregate to logical plan"),

src/daft-connect/src/translation/logical_plan/filter.rs

@@ -0,0 +1,22 @@
+use eyre::bail;
+
+use crate::translation::{to_daft_expr, to_logical_plan, Plan};
+
+pub async fn filter(filter: spark_connect::Filter) -> eyre::Result<Plan> {
+    let spark_connect::Filter { input, condition } = filter;
+
+    let Some(input) = input else {
+        bail!("input is required");
+    };
+
+    let Some(condition) = condition else {
+        bail!("condition is required");
+    };
+
+    let condition = to_daft_expr(&condition)?;
+
+    let mut plan = Box::pin(to_logical_plan(*input)).await?;
+    plan.builder = plan.builder.filter(condition)?;
+
+    Ok(plan)
+}

tests/connect/test_filter.py

@@ -0,0 +1,19 @@
+from __future__ import annotations
+
+from pyspark.sql.functions import col
+
+
+def test_filter(spark_session):
+    # Create DataFrame from range(10)
+    df = spark_session.range(10)
+
+    # Filter for values less than 5
+    df_filtered = df.filter(col("id") < 5)
+
+    # Verify the schema is unchanged after filter
+    assert df_filtered.schema == df.schema, "Schema should be unchanged after filter"
+
+    # Verify the filtered data is correct
+    df_filtered_pandas = df_filtered.toPandas()
+    assert len(df_filtered_pandas) == 5, "Should have 5 rows after filtering < 5"
+    assert all(df_filtered_pandas["id"] < 5), "All values should be less than 5"