waynexia commented on code in PR #2587: URL: https://github.com/apache/arrow-datafusion/pull/2587#discussion_r880063760
########## datafusion/jit/src/compile.rs: ########## @@ -0,0 +1,184 @@ +// 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. + +//! Compile DataFusion Expr to JIT'd function. + +use datafusion_common::Result; + +use crate::api::Assembler; +use crate::{ + api::GeneratedFunction, + ast::{Expr as JITExpr, I64, PTR_SIZE}, +}; + +/// Wrap JIT Expr to array compute function. +pub fn build_calc_fn( + assembler: &Assembler, + jit_expr: JITExpr, + input_names: Vec<String>, +) -> Result<GeneratedFunction> { + let mut builder = assembler.new_func_builder("calc_fn"); + for input in &input_names { + builder = builder.param(format!("{}_array", input), I64); + } + let mut builder = builder.param("result", I64).param("len", I64); + + let mut fn_body = builder.enter_block(); + fn_body.declare_as("index", fn_body.lit_i(0))?; + fn_body.while_block( + |cond| cond.lt(cond.id("index")?, cond.id("len")?), + |w| { + w.declare_as("offset", w.mul(w.id("index")?, w.lit_i(PTR_SIZE as i64))?)?; + for input in &input_names { + w.declare_as( + format!("{}_ptr", input), + w.add(w.id(format!("{}_array", input))?, w.id("offset")?)?, + )?; + w.declare_as(input, w.deref(w.id(format!("{}_ptr", input))?, I64)?)?; + } + w.declare_as("res_ptr", w.add(w.id("result")?, w.id("offset")?)?)?; + w.declare_as("res", jit_expr.clone())?; + w.store(w.id("res")?, w.id("res_ptr")?)?; + + w.assign("index", w.add(w.id("index")?, w.lit_i(1))?)?; + Ok(()) + }, + )?; + + let gen_func = fn_body.build(); + Ok(gen_func) +} + +#[cfg(test)] +mod test { + use std::{collections::HashMap, sync::Arc}; + + use arrow::{ + array::{Array, PrimitiveArray}, + datatypes::{DataType, Int64Type}, + }; + use datafusion_common::{DFField, DFSchema, DataFusionError}; + use datafusion_expr::Expr as DFExpr; + + use crate::ast::BinaryExpr; + + use super::*; + + fn run_df_expr( Review Comment: Definitely! I'm going to figure out how to play with var length param list and type casting, and then encapsulate a plan node level interface (maybe take `LogicalPlan` and data as input, haven't thought in detail) as next step. -- 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]
