Raise IndexError and KeyError for invalid column accessors

.github/workflows/docs.yml

@@ -45,9 +45,9 @@ jobs:
       - name: Build python packages
         run: |
           # arro3-core needs to be first
-          uv run maturin develop -m arro3-core/Cargo.toml
-          uv run maturin develop -m arro3-compute/Cargo.toml
-          uv run maturin develop -m arro3-io/Cargo.toml
+          uv run maturin dev -m arro3-core/Cargo.toml
+          uv run maturin dev -m arro3-compute/Cargo.toml
+          uv run maturin dev -m arro3-io/Cargo.toml
 
       - name: Deploy docs
         env:

.github/workflows/test-python.yml

@@ -62,9 +62,9 @@ jobs:
       - name: Build rust submodules
         run: |
           # Note: core module must be first, because it's depended on by others
-          uv run maturin develop -m arro3-core/Cargo.toml
-          uv run maturin develop -m arro3-compute/Cargo.toml
-          uv run maturin develop -m arro3-io/Cargo.toml
+          uv run maturin dev -m arro3-core/Cargo.toml
+          uv run maturin dev -m arro3-compute/Cargo.toml
+          uv run maturin dev -m arro3-io/Cargo.toml
 
       - name: Run python tests
         run: |

CHANGELOG.md

@@ -2,6 +2,12 @@
 
 This is the changelog for arro3. pyo3-arrow has a separate changelog.
 
+## [0.4.4] - 2024-12-09
+
+### Bug fixes :bug:
+
+- Raise IndexError and KeyError for invalid column access https://github.com/kylebarron/arro3/pull/272
+
 ## [0.4.3] - 2024-11-21
 
 ### What's Changed

Cargo.toml

@@ -9,7 +9,7 @@ resolver = "2"
 
 [workspace.package]
 # Package version for arro3-*, not for pyo3-arrow
-version = "0.4.3"
+version = "0.4.4"
 authors = ["Kyle Barron <[email protected]>"]
 edition = "2021"
 homepage = "https://kylebarron.dev/arro3"

DEVELOP.md

@@ -4,9 +4,9 @@
 rm -rf .venv
 uv sync
 # Note: need to install core first because others depend on core
-uv run maturin develop -m arro3-core/Cargo.toml
-uv run maturin develop -m arro3-compute/Cargo.toml
-uv run maturin develop -m arro3-io/Cargo.toml
+uv run maturin dev -m arro3-core/Cargo.toml
+uv run maturin dev -m arro3-compute/Cargo.toml
+uv run maturin dev -m arro3-io/Cargo.toml
 uv run mkdocs serve
 ```
 

pyo3-arrow/src/input.rs

@@ -9,7 +9,7 @@ use std::sync::Arc;
 
 use arrow_array::{Datum, RecordBatchIterator, RecordBatchReader};
 use arrow_schema::{ArrowError, Field, FieldRef, Fields, Schema, SchemaRef};
-use pyo3::exceptions::PyValueError;
+use pyo3::exceptions::{PyIndexError, PyKeyError, PyValueError};
 use pyo3::prelude::*;
 
 use crate::array_reader::PyArrayReader;
@@ -162,10 +162,21 @@ pub(crate) enum FieldIndexInput {
 }
 
 impl FieldIndexInput {
-    pub fn into_position(self, schema: &Schema) -> PyArrowResult<usize> {
+    /// This will additionally check that the input is valid against the given schema.
+    ///
+    /// This will raise a KeyError if the provided name does not exist, or an IndexError if the
+    /// provided integer index is out of bounds.
+    pub fn into_position(self, schema: &Schema) -> PyResult<usize> {
         match self {
-            Self::Name(name) => Ok(schema.index_of(name.as_ref())?),
-            Self::Position(position) => Ok(position),
+            Self::Name(name) => schema
+                .index_of(name.as_ref())
+                .map_err(|err| PyKeyError::new_err(err.to_string())),
+            Self::Position(position) => {
+                if position >= schema.fields().len() {
+                    return Err(PyIndexError::new_err("Index out of range").into());
+                }
+                Ok(position)
+            }
         }
     }
 }

tests/core/test_table.py

@@ -10,11 +10,18 @@ def test_table_getitem():
     a = pa.chunked_array([[1, 2, 3, 4]])
     b = pa.chunked_array([["a", "b", "c", "d"]])
     table = Table.from_pydict({"a": a, "b": b})
+
     assert a == pa.chunked_array(table["a"])
     assert b == pa.chunked_array(table["b"])
     assert a == pa.chunked_array(table[0])
     assert b == pa.chunked_array(table[1])
 
+    with pytest.raises(KeyError):
+        table["foo"]
+
+    with pytest.raises(IndexError):
+        table[10]
+
 
 def test_table_from_arrays():
     a = pa.array([1, 2, 3, 4])