LiaCastaneda commented on code in PR #18921: URL: https://github.com/apache/datafusion/pull/18921#discussion_r2996271140
########## datafusion/functions-nested/src/array_transform.rs: ########## @@ -0,0 +1,253 @@ +// Licensed to the Apache Software Foundation (ASF) under one +// or more contributor license agreements. See the NOTICE file +// distributed with this work for additional information +// regarding copyright ownership. The ASF licenses this file +// to you under the Apache License, Version 2.0 (the +// "License"); you may not use this file except in compliance +// with the License. You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, +// software distributed under the License is distributed on an +// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +// KIND, either express or implied. See the License for the +// specific language governing permissions and limitations +// under the License. + +//! [`LambdaUDF`] definitions for array_transform function. + +use arrow::{ + array::{Array, ArrayRef, AsArray, FixedSizeListArray, LargeListArray, ListArray}, + datatypes::{DataType, Field, FieldRef}, +}; +use datafusion_common::{ + Result, exec_err, plan_err, + utils::{list_values, take_function_args}, +}; +use datafusion_expr::{ + ColumnarValue, Documentation, LambdaFunctionArgs, LambdaReturnFieldArgs, + LambdaSignature, LambdaUDF, ValueOrLambda, Volatility, +}; +use datafusion_macros::user_doc; +use std::{any::Any, fmt::Debug, sync::Arc}; + +make_udlf_expr_and_func!( + ArrayTransform, + array_transform, + array lambda, + "transforms the values of a array", + array_transform_udlf +); + +#[user_doc( + doc_section(label = "Array Functions"), + description = "transforms the values of a array", + syntax_example = "array_transform(array, x -> x*2)", + sql_example = r#"```sql +> select array_transform([1, 2, 3, 4, 5], x -> x*2); ++-------------------------------------------+ +| array_transform([1, 2, 3, 4, 5], x -> x*2) | ++-------------------------------------------+ +| [2, 4, 6, 8, 10] | ++-------------------------------------------+ +```"#, + argument( + name = "array", + description = "Array expression. Can be a constant, column, or function, and any combination of array operators." + ), + argument(name = "lambda", description = "Lambda") +)] +#[derive(Debug, PartialEq, Eq, Hash)] +pub struct ArrayTransform { + signature: LambdaSignature, + aliases: Vec<String>, +} + +impl Default for ArrayTransform { + fn default() -> Self { + Self::new() + } +} + +impl ArrayTransform { + pub fn new() -> Self { + Self { + signature: LambdaSignature::user_defined(Volatility::Immutable), + aliases: vec![String::from("list_transform")], + } + } +} + +impl LambdaUDF for ArrayTransform { + fn as_any(&self) -> &dyn Any { + self + } + + fn name(&self) -> &str { + "array_transform" + } + + fn aliases(&self) -> &[String] { + &self.aliases + } + + fn signature(&self) -> &LambdaSignature { + &self.signature + } + + fn coerce_value_types(&self, arg_types: &[DataType]) -> Result<Vec<DataType>> { + let list = if arg_types.len() == 1 { + &arg_types[0] + } else { + return plan_err!( + "{} function requires 1 value arguments, got {}", + self.name(), + arg_types.len() + ); + }; + + let coerced = match list { + DataType::List(_) + | DataType::LargeList(_) + | DataType::FixedSizeList(_, _) => list.clone(), + DataType::ListView(field) => DataType::List(Arc::clone(field)), + DataType::LargeListView(field) => DataType::LargeList(Arc::clone(field)), + _ => { + return plan_err!( + "{} expected a list as first argument, got {}", + self.name(), + list + ); + } + }; + + Ok(vec![coerced]) + } + + fn lambdas_parameters( + &self, + args: &[ValueOrLambda<FieldRef, ()>], + ) -> Result<Vec<Option<Vec<Field>>>> { Review Comment: > I kept value_fields_with_lambda_udf working with the ValueOrLambda enum because it needs to check the number of all arguments for LambdaTypeSignature::Any(usize) and if/when we add TypeSignature::Exact equivalent it will likely define the exact position of lambdas, which will require full positional info yeah I can't think of another way of removing `ValueOrLambda` from `value_fields_with_lambda_udf` and still verify that we have the correct positional info. > Just one minor thing to note is that both changes at coerce_value_types and lambas_parameters postpone the detection of wrong parameters positions to return_field_from_args, which it's not ideal but it's not a big deal either An error would happen regardless, just caught by Df internally rather than the implementor right? imo I think it's fine since validating argument ordering is a bit out of scope for `coerce_value_types` or `lambas_parameters` anyway. If we see a lambda function in the future that needs strict position validation, we could expose a method in the trait for that. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
