jvanstraten commented on a change in pull request #12279: URL: https://github.com/apache/arrow/pull/12279#discussion_r798572833
########## File path: format/substrait/extension_types.yaml ########## @@ -0,0 +1,121 @@ +# 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. + +# substrait::{ExtensionTypeVariation, ExtensionType}s +# for wrapping types which appear in the arrow type system but +# are not first-class in substrait. These include: +# - null +# - unsigned integers +# - half-precision floating point numbers +# - 32 bit times and dates +# - timestamps with units other than microseconds +# - timestamps with timezones other than UTC +# - 256 bit decimals +# - sparse and dense unions +# - dictionary encoded types +# - durations +# - string and binary with 64 bit offsets +# - list with 64 bit offsets +# - interval<months: i32> +# - interval<days: i32, millis: i32> +# - interval<months: i32, days: i32, nanos: i64> +# - arrow::ExtensionTypes + +# FIXME these extension types are not parameterizable, which means among +# other things that we can't declare dictionary type here at all since +# we'd have to declare a different dictionary type for all encoded types +# (but that is an infinite space). Similarly, do we need to declare a +# timestamp variation for all possible timezone strings? +# +# Ultimately these declarations are a promise which needs to be backed by +# equivalent serde in c++. For example, consider u8: when serializing to +# substrait, we need to wrap instances of arrow::uint8 into the type +# variation listed below. It would be ideal if we could SinglePointOfTruth +# this correspondence; either generating c++ from the YAML or YAML from the +# c++. Review comment: I basically replaced that sentence with info on which two files need to be kept in sync right now (and also in the reverse direction, which I do think is important for anyone who wants to add a mapping) and a reference to the issue tracking its automation in cbc9842. Let me know if that works for you. -- 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: github-unsubscr...@arrow.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
