bkietz commented on a change in pull request #10934: URL: https://github.com/apache/arrow/pull/10934#discussion_r689823790
########## File path: format/ComputeIR.fbs ########## @@ -0,0 +1,241 @@ +// 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. + +include "Schema.fbs"; +include "Message.fbs"; + +namespace org.apache.arrow.flatbuf.computeir; + +/// An expression is one of +/// - a Literal datum +/// - a reference to a Field from a Relation +/// - a call to a named function +/// On evaluation, an Expression will have either array or scalar shape. +union ExpressionImpl { + Literal, FieldRef, Call +} + +table Expression { + impl: ExpressionImpl (required); +} + +union Shape { + ArrayShape, ScalarShape +} + +table ScalarShape {} + +table ArrayShape { + /// Number of slots. + length: long; +} + +table Literal { + /// Shape of this literal. + shape: Shape (required); + + /// The type of this literal. + type: Type (required); + + /// Buffers containing `length` elements of arrow-formatted data. + /// If `length` is absent (this Literal is scalar), these buffers + /// are sized to accommodate a single element of arrow-formatted data. + /// XXX this can be optimized for trivial scalars later + buffers: [Buffer]; +} + +table FieldRef { + /// A sequence of field names to allow referencing potentially nested fields + path: [string]; + + /// For Expressions which might reference fields in multiple Relations, + /// this index may be provided to indicate which Relation's fields + /// `path` points into. For example in the case of a join, + /// 0 refers to the left relation and 1 to the right relation. + relation_index: int; + + /// The type of data in the referenced Field. + type: Type; +} + +table Call { + /// The name of the function whose invocation this Call represents. + function_name: string (required); + + /// Parameters for `function_name`; content/format may be unique to each + /// value of `function_name`. + options: Buffer; Review comment: Yes, they are defined as ```c++ /// A Buffer represents a single contiguous memory segment struct Buffer { /// The relative offset into the shared memory page where the bytes for this /// buffer starts offset: long; /// The absolute length (in bytes) of the memory buffer. The memory is found /// from offset (inclusive) to offset + length (non-inclusive). When building /// messages using the encapsulated IPC message, padding bytes may be written /// after a buffer, but such padding bytes do not need to be accounted for in /// the size here. length: long; } ``` IE they reference memory which is out-of-band WRT flatbuffers. This gives producers of Buffers control over padding and alignment, which `[ubyte]` does not -- 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]
