Function write_int_parquet_file

crates/test_utils.rs:29–55

(
    path: &Path,
    columns: Vec<(&str, Vec<i32>)>,
    max_row_group_size: Option<usize>,
)

use serde::de::DeserializeOwned;

pub(crate) fn write_int_parquet_file(
    path: &Path,
    columns: Vec<(&str, Vec<i32>)>,
    max_row_group_size: Option<usize>,
) {
    let schema = Arc::new(ArrowSchema::new(
        columns
            .iter()
            .map(|(name, _)| ArrowField::new(*name, ArrowDataType::Int32, false))
            .collect::<Vec<_>>(),
    ));
    let arrays: Vec<Arc<dyn Array>> = columns
        .iter()
        .map(|(_, values)| Arc::new(Int32Array::from(values.clone())) as Arc<dyn Array>)
        .collect();
    let batch = RecordBatch::try_new(schema.clone(), arrays).unwrap();

    let props = max_row_group_size.map(|size| {
        WriterProperties::builder()
            .set_max_row_group_row_count(Some(size))
            .build()
    });
    let file = File::create(path).unwrap();
    let mut writer = ArrowWriter::try_new(file, schema, props).unwrap();
    writer.write(&batch).unwrap();
    writer.close().unwrap();
}

pub(crate) fn local_file_path(path: &Path) -> String {
    let normalized = path.to_string_lossy().replace('\\', "/");