forked from cockroachdb/cockroach
-
Notifications
You must be signed in to change notification settings - Fork 0
/
generator_builtins.go
240 lines (211 loc) · 7.16 KB
/
generator_builtins.go
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
// Copyright 2016 The Cockroach Authors.
//
// Licensed 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.
//
// Author: Raphael 'kena' Poss (knz@cockroachlabs.com)
package parser
import (
"errors"
"fmt"
)
// Table generators, also called "set-generating functions", are
// special functions that return an entire table.
//
// Overview of the concepts:
//
// - generators are implemented as regular built-in functions that
// return values of type DTable (this is a Datum type).
//
// - the return type of generators is a TTable. This describes objects
// that are conceptually sets of rows. A TTable type is
// characterized by its column types (no names).
//
// - a DTable doesn't carry the contents of a table directly; instead
// it carries a ValueGenerator reference.
//
// - ValueGenerator is an interface that offers a
// Start/Next/Values/Stop API similar to sql.planNode.
//
// - because generators are regular functions, it is possible to use
// them in any expression context. This is useful to e.g
// pass an entire table as argument to the ARRAY( ) conversion
// function.
//
// - the data source mechanism in the sql package has a special case
// for generators appearing in FROM contexts and knows how to
// construct a special row source from them (the valueGenerator
// planNode).
// generatorFactory is the type of constructor functions for
// ValueGenerator objects suitable for use with DTable.
type generatorFactory func(ctx *EvalContext, args Datums) (ValueGenerator, error)
// ValueGenerator is the interface provided by the object held by a
// DTable; objects that implement this interface are able to produce
// rows of values in a streaming fashion (like Go iterators or
// generators in Python).
type ValueGenerator interface {
// ColumnTypes returns the type signature of this value generator.
// Used by DTable.ResolvedType().
ColumnTypes() TTuple
// Start initializes the generator. Must be called once before
// Next() and Values().
Start() error
// Next determines whether there is a row of data available.
Next() (bool, error)
// Values retrieves the current row of data.
Values() Datums
// Close must be called after Start() before disposing of the
// ValueGenerator. It does not need to be called if Start() has not
// been called yet.
Close()
}
var _ ValueGenerator = &seriesValueGenerator{}
var _ ValueGenerator = &arrayValueGenerator{}
func initGeneratorBuiltins() {
// Add all windows to the Builtins map after a few sanity checks.
for k, v := range Generators {
for _, g := range v {
if !g.impure {
panic(fmt.Sprintf("generator functions should all be impure, found %v", g))
}
if g.class != GeneratorClass {
panic(fmt.Sprintf("generator functions should be marked with the GeneratorClass "+
"function class, found %v", g))
}
}
Builtins[k] = v
}
}
// Generators is a map from name to slice of Builtins for all built-in
// generators.
var Generators = map[string][]Builtin{
"generate_series": {
makeGeneratorBuiltin(
ArgTypes{{"start", TypeInt}, {"end", TypeInt}},
TTuple{TypeInt},
makeSeriesGenerator,
"Produces a virtual table containing the integer values from `start` to `end`, inclusive.",
),
makeGeneratorBuiltin(
ArgTypes{{"start", TypeInt}, {"end", TypeInt}, {"step", TypeInt}},
TTuple{TypeInt},
makeSeriesGenerator,
"Produces a virtual table containing the integer values from `start` to `end`, inclusive, by increment of `step`.",
),
},
"unnest": {
makeGeneratorBuiltinWithReturnType(
ArgTypes{{"input", TypeAnyArray}},
func(args []TypedExpr) Type {
if len(args) == 0 {
return unknownReturnType
}
return TTable{Cols: TTuple{args[0].ResolvedType().(TArray).Typ}}
},
makeArrayGenerator,
"Returns the input array as a set of rows",
),
},
}
func makeGeneratorBuiltin(in ArgTypes, ret TTuple, g generatorFactory, info string) Builtin {
return makeGeneratorBuiltinWithReturnType(in, fixedReturnType(TTable{Cols: ret}), g, info)
}
func makeGeneratorBuiltinWithReturnType(
in ArgTypes, retType returnTyper, g generatorFactory, info string,
) Builtin {
return Builtin{
impure: true,
class: GeneratorClass,
Types: in,
ReturnType: retType,
fn: func(ctx *EvalContext, args Datums) (Datum, error) {
gen, err := g(ctx, args)
if err != nil {
return nil, err
}
return &DTable{gen}, nil
},
category: categoryCompatibility,
Info: info,
}
}
// seriesValueGenerator supports the execution of generate_series()
// with integer bounds.
type seriesValueGenerator struct {
value, start, stop, step int64
}
var errStepCannotBeZero = errors.New("step cannot be 0")
func makeSeriesGenerator(_ *EvalContext, args Datums) (ValueGenerator, error) {
start := int64(MustBeDInt(args[0]))
stop := int64(MustBeDInt(args[1]))
step := int64(1)
if len(args) > 2 {
step = int64(MustBeDInt(args[2]))
}
if step == 0 {
return nil, errStepCannotBeZero
}
return &seriesValueGenerator{start, start, stop, step}, nil
}
// ColumnTypes implements the ValueGenerator interface.
func (s *seriesValueGenerator) ColumnTypes() TTuple { return TTuple{TypeInt} }
// Start implements the ValueGenerator interface.
func (s *seriesValueGenerator) Start() error { return nil }
// Close implements the ValueGenerator interface.
func (s *seriesValueGenerator) Close() {}
// Next implements the ValueGenerator interface.
func (s *seriesValueGenerator) Next() (bool, error) {
if s.step < 0 && (s.start < s.stop) {
return false, nil
}
if s.step > 0 && (s.stop < s.start) {
return false, nil
}
s.value = s.start
s.start += s.step
return true, nil
}
// Values implements the ValueGenerator interface.
func (s *seriesValueGenerator) Values() Datums {
return Datums{NewDInt(DInt(s.value))}
}
func makeArrayGenerator(_ *EvalContext, args Datums) (ValueGenerator, error) {
arr := MustBeDArray(args[0])
return &arrayValueGenerator{array: arr}, nil
}
// arrayValueGenerator is a value generator that returns each element of an
// array.
type arrayValueGenerator struct {
array *DArray
nextIndex int
}
// ColumnTypes implements the ValueGenerator interface.
func (s *arrayValueGenerator) ColumnTypes() TTuple { return TTuple{s.array.ParamTyp} }
// Start implements the ValueGenerator interface.
func (s *arrayValueGenerator) Start() error {
s.nextIndex = -1
return nil
}
// Close implements the ValueGenerator interface.
func (s *arrayValueGenerator) Close() {}
// Next implements the ValueGenerator interface.
func (s *arrayValueGenerator) Next() (bool, error) {
s.nextIndex++
if s.nextIndex >= s.array.Len() {
return false, nil
}
return true, nil
}
// Values implements the ValueGenerator interface.
func (s *arrayValueGenerator) Values() Datums {
return Datums{s.array.Array[s.nextIndex]}
}