new operator 'groupby'

This commit is contained in:
Celestino Amoroso 2026-05-02 14:46:28 +02:00
parent d5ced343c4
commit 3b2ef7927b
12 changed files with 271 additions and 58 deletions

View File

@ -143,13 +143,12 @@ func (dict *DictType) HasKey(target any) (ok bool) {
return return
} }
func (dict *DictType) SetItem(key any, value any) (err error) { func (dict *DictType) SetItem(key any, value any) {
(*dict)[key] = value (*dict)[key] = value
return
} }
func (dict *DictType) GetItem(key any) (value any, err error) { func (dict *DictType) GetItem(key any) (value any, exists bool) {
value = (*dict)[key] value, exists = (*dict)[key]
return return
} }
@ -169,6 +168,24 @@ func (dict *DictType) Merge(second *DictType) {
} }
} }
func (dict *DictType) Equals(dict2 DictType) (answer bool) {
if dict2 != nil && len(*dict) == len(dict2) {
answer = true
for key, value1 := range *dict {
if value2, exists := dict2.GetItem(key); exists {
if !Equal(value1, value2) {
answer = false
break
}
} else {
answer = false
break
}
}
}
return
}
//////////////// ////////////////
type DictFormat interface { type DictFormat interface {

View File

@ -220,11 +220,12 @@ func anyToFract(v any) (f *FractionType, err error) {
if f, ok = v.(*FractionType); !ok { if f, ok = v.(*FractionType); !ok {
if n, ok := v.(int64); ok { if n, ok := v.(int64); ok {
f = intToFraction(n) f = intToFraction(n)
} else if dec, ok := v.(float64); ok {
f, err = Float64ToFraction(dec)
} else {
err = ErrExpectedGot("fract", TypeFraction, v)
} }
} }
if f == nil {
err = ErrExpectedGot("fract", TypeFraction, v)
}
return return
} }
@ -273,14 +274,16 @@ func CmpAnyFract(af1, af2 any) (result int, err error) {
// =0 if af1 == af2 // =0 if af1 == af2
// >0 if af1 > af2 // >0 if af1 > af2
func cmpFract(f1, f2 *FractionType) (result int) { func cmpFract(f1, f2 *FractionType) (result int) {
f2.num = -f2.num if f1 != nil && f2 != nil {
f := SumFract(f1, f2) f2.num = -f2.num
if f.num < 0 { f := SumFract(f1, f2)
result = -1 if f.num < 0 {
} else if f.num > 0 { result = -1
result = 1 } else if f.num > 0 {
} else { result = 1
result = 0 } else {
result = 0
}
} }
return return
} }

View File

@ -50,13 +50,13 @@ func ListFromStrings(stringList []string) (list *ListType) {
return return
} }
func (ls *ListType) ToString(opt FmtOpt) (s string) { func (dict *ListType) ToString(opt FmtOpt) (s string) {
indent := GetFormatIndent(opt) indent := GetFormatIndent(opt)
flags := GetFormatFlags(opt) flags := GetFormatFlags(opt)
var sb strings.Builder var sb strings.Builder
sb.WriteByte('[') sb.WriteByte('[')
if len(*ls) > 0 { if len(*dict) > 0 {
innerOpt := MakeFormatOptions(flags, indent+1) innerOpt := MakeFormatOptions(flags, indent+1)
nest := strings.Repeat(" ", indent+1) nest := strings.Repeat(" ", indent+1)
@ -64,7 +64,7 @@ func (ls *ListType) ToString(opt FmtOpt) (s string) {
sb.WriteByte('\n') sb.WriteByte('\n')
sb.WriteString(nest) sb.WriteString(nest)
} }
for i, item := range []any(*ls) { for i, item := range []any(*dict) {
if i > 0 { if i > 0 {
if flags&MultiLine != 0 { if flags&MultiLine != 0 {
sb.WriteString(",\n") sb.WriteString(",\n")
@ -96,19 +96,19 @@ func (ls *ListType) ToString(opt FmtOpt) (s string) {
return return
} }
func (ls *ListType) String() string { func (dict *ListType) String() string {
return ls.ToString(0) return dict.ToString(0)
} }
func (ls *ListType) TypeName() string { func (dict *ListType) TypeName() string {
return "list" return "list"
} }
func (ls *ListType) Contains(t *ListType) (answer bool) { func (dict *ListType) Contains(t *ListType) (answer bool) {
if len(*ls) >= len(*t) { if len(*dict) >= len(*t) {
answer = true answer = true
for _, item := range *t { for _, item := range *t {
if answer = ls.IndexDeepSameCmp(item) >= 0; !answer { if answer = dict.IndexDeepSameCmp(item) >= 0; !answer {
break break
} }
} }
@ -120,8 +120,11 @@ func (ls1 *ListType) Equals(ls2 ListType) (answer bool) {
if ls2 != nil && len(*ls1) == len(ls2) { if ls2 != nil && len(*ls1) == len(ls2) {
answer = true answer = true
for index, i1 := range *ls1 { for index, i1 := range *ls1 {
// if i1 != (ls2)[index] { // if !reflect.DeepEqual(i1, ls2[index]) {
if !reflect.DeepEqual(i1, ls2[index]) { // answer = false
// break
// }
if !Equal(i1, ls2[index]) {
answer = false answer = false
break break
} }
@ -130,11 +133,11 @@ func (ls1 *ListType) Equals(ls2 ListType) (answer bool) {
return return
} }
func (list *ListType) IndexDeepSameCmp(target any) (index int) { func (dict *ListType) IndexDeepSameCmp(target any) (index int) {
var eq bool var eq bool
var err error var err error
index = -1 index = -1
for i, item := range *list { for i, item := range *dict {
if eq, err = deepSame(item, target, SameContent); err != nil { if eq, err = deepSame(item, target, SameContent); err != nil {
break break
} else if eq { } else if eq {
@ -185,15 +188,15 @@ func deepSame(a, b any, deepCmp DeepFuncTemplate) (eq bool, err error) {
return return
} }
func (list *ListType) SetItem(index int64, value any) (err error) { func (dict *ListType) SetItem(index int64, value any) (err error) {
if index >= 0 && index < int64(len(*list)) { if index >= 0 && index < int64(len(*dict)) {
(*list)[index] = value (*dict)[index] = value
} else { } else {
err = fmt.Errorf("index %d out of bounds (0, %d)", index, len(*list)-1) err = fmt.Errorf("index %d out of bounds (0, %d)", index, len(*dict)-1)
} }
return return
} }
func (list *ListType) AppendItem(value any) { func (dict *ListType) AppendItem(value any) {
*list = append(*list, value) *dict = append(*dict, value)
} }

View File

@ -49,7 +49,7 @@ func assignCollectionItem(ctx kern.ExprContext, collectionTerm, keyListTerm *ter
err = keyListTerm.Errorf("integer expected, got %v [%s]", keyValue, kern.TypeName(keyValue)) err = keyListTerm.Errorf("integer expected, got %v [%s]", keyValue, kern.TypeName(keyValue))
} }
case *kern.DictType: case *kern.DictType:
err = collection.SetItem(keyValue, value) collection.SetItem(keyValue, value)
default: default:
err = collectionTerm.Errorf("collection expected") err = collectionTerm.Errorf("collection expected")
} }

View File

@ -45,15 +45,20 @@ func evalDot(ctx kern.ExprContext, opTerm *term) (v any, err error) {
err = indexTerm.tk.ErrorExpectedGot("identifier") err = indexTerm.tk.ErrorExpectedGot("identifier")
} }
case *kern.DictType: case *kern.DictType:
var ok bool
s := opTerm.children[1].symbol() s := opTerm.children[1].symbol()
if s == SymVariable || s == SymString { if s == SymVariable || s == SymString {
src := opTerm.children[1].Source() src := opTerm.children[1].Source()
if len(src) > 1 && src[0] == '"' && src[len(src)-1] == '"' { if len(src) > 1 && src[0] == '"' && src[len(src)-1] == '"' {
src = src[1 : len(src)-1] src = src[1 : len(src)-1]
} }
v, err = unboxedValue.GetItem(src) if v, ok = unboxedValue.GetItem(src); !ok {
err = opTerm.errKeyNotFound(src)
}
} else if rightValue, err = opTerm.children[1].Compute(ctx); err == nil { } else if rightValue, err = opTerm.children[1].Compute(ctx); err == nil {
v, err = unboxedValue.GetItem(rightValue) if v, ok = unboxedValue.GetItem(rightValue); !ok {
err = opTerm.errKeyNotFound(rightValue)
}
} }
default: default:
if rightValue, err = opTerm.children[1].Compute(ctx); err == nil { if rightValue, err = opTerm.children[1].Compute(ctx); err == nil {

105
operator-groupby.go Normal file
View File

@ -0,0 +1,105 @@
// Copyright (c) 2024-2026 Celestino Amoroso (celestino.amoroso@gmail.com).
// All rights reserved.
// operator-groupby.go
package expr
import (
"fmt"
"io"
"strconv"
"git.portale-stac.it/go-pkg/expr/kern"
)
//-------- group by term
func newGroupByTerm(tk *Token) (inst *term) {
return &term{
tk: *tk,
children: make([]*term, 0, 2),
position: posInfix,
priority: priIterOp,
evalFunc: evalGroupBy,
}
}
func evalGroupBy(ctx kern.ExprContext, opTerm *term) (v any, err error) {
var leftValue, rightValue any
var it kern.Iterator
var item any
var sKey string
var keyByIndex bool
if err = opTerm.checkOperands(); err != nil {
return
}
if leftValue, err = opTerm.children[0].Compute(ctx); err != nil {
return
}
if it, err = NewIterator(leftValue); err != nil {
return nil, fmt.Errorf("left operand of MAP must be an iterable data-source; got %s", kern.TypeName(leftValue))
}
rightTk := opTerm.children[1].tk
if rightTk.IsSymbol(SymVariable) && rightTk.source == "__" {
keyByIndex = true
} else if rightValue, err = opTerm.children[1].Compute(ctx); err != nil {
return
} else if kern.IsString(rightValue) {
sKey = rightValue.(string)
} else {
return nil, fmt.Errorf("right operand of GROUPBY must be a string or identifier '__'; got %s", kern.TypeName(rightValue))
}
values := kern.MakeDict()
for item, err = it.Next(); err == nil; item, err = it.Next() {
ctx.SetVar("_", item)
ctx.SetVar("__", it.Index())
ctx.SetVar("_#", it.Count())
var sItemKey string
if d, ok := item.(*kern.DictType); ok {
if keyByIndex || len(sKey) == 0 {
sItemKey = strconv.Itoa(int(it.Index()))
} else if d.HasKey(sKey) {
if keyValue, exists := d.GetItem(sKey); exists {
sItemKey = fmt.Sprintf("%v", keyValue)
} else {
sItemKey = "_"
}
} else {
sItemKey = "_"
}
} else {
sItemKey = strconv.Itoa(int(it.Index()))
}
var ls *kern.ListType
if lsAny, exists := values.GetItem(sItemKey); exists && lsAny != nil {
ls = lsAny.(*kern.ListType)
}
if ls == nil {
ls = kern.NewListA()
}
ls.AppendItem(item)
values.SetItem(sItemKey, ls)
ctx.DeleteVar("_#")
ctx.DeleteVar("__")
ctx.DeleteVar("_")
}
if err == io.EOF {
err = nil
}
v = values
return
}
// init
func init() {
registerTermConstructor(SymKwGroupBy, newGroupByTerm)
}

View File

@ -139,6 +139,7 @@ func init() {
SymKwFilter: {"filter", symClassOperator, posInfix}, SymKwFilter: {"filter", symClassOperator, posInfix},
SymKwDigest: {"digest", symClassOperator, posInfix}, SymKwDigest: {"digest", symClassOperator, posInfix},
SymKwJoin: {"join", symClassOperator, posInfix}, SymKwJoin: {"join", symClassOperator, posInfix},
SymKwGroupBy: {"groupby", symClassOperator, posInfix},
SymKwFunc: {"func(", symClassDeclaration, posPrefix}, SymKwFunc: {"func(", symClassDeclaration, posPrefix},
SymKwBuiltin: {"builtin", symClassOperator, posPrefix}, SymKwBuiltin: {"builtin", symClassOperator, posPrefix},
SymKwPlugin: {"plugin", symClassOperator, posPrefix}, SymKwPlugin: {"plugin", symClassOperator, posPrefix},

View File

@ -122,6 +122,7 @@ const (
SymKwMap SymKwMap
SymKwFilter SymKwFilter
SymKwDigest SymKwDigest
SymKwGroupBy
SymKwJoin SymKwJoin
SymKwNil SymKwNil
SymKwUnset SymKwUnset
@ -147,5 +148,6 @@ func init() {
"UNSET": SymKwUnset, "UNSET": SymKwUnset,
"DIGEST": SymKwDigest, "DIGEST": SymKwDigest,
"JOIN": SymKwJoin, "JOIN": SymKwJoin,
"GROUPBY": SymKwGroupBy,
} }
} }

View File

@ -6,7 +6,6 @@ package expr
import ( import (
"errors" "errors"
"reflect"
"strings" "strings"
"testing" "testing"
@ -74,7 +73,6 @@ func doTest(t *testing.T, ctx kern.ExprContext, section string, input *inputType
var ast Expr var ast Expr
var gotResult any var gotResult any
var gotErr error var gotErr error
var eq, eqDone bool
wantErr := getWantedError(input) wantErr := getWantedError(input)
@ -93,18 +91,24 @@ func doTest(t *testing.T, ctx kern.ExprContext, section string, input *inputType
gotResult, gotErr = ast.Eval(ctx) gotResult, gotErr = ast.Eval(ctx)
} }
if input.wantResult != nil && gotResult != nil { eq := kern.Equal(gotResult, input.wantResult)
if ls1, ok := input.wantResult.(*kern.ListType); ok { // if input.wantResult != nil && gotResult != nil {
if ls2, ok := gotResult.(*kern.ListType); ok { // if ls1, ok := input.wantResult.(*kern.ListType); ok {
eq = ls1.Equals(*ls2) // if ls2, ok := gotResult.(*kern.ListType); ok {
eqDone = true // eq = ls1.Equals(*ls2)
} // eqDone = true
} // }
} // } else if dict1, ok := input.wantResult.(*kern.DictType); ok {
// if dict2, ok := gotResult.(*kern.DictType); ok {
// eq = dict1.Equals(*dict2)
// eqDone = true
// }
// }
// }
if !eqDone { // if !eqDone {
eq = reflect.DeepEqual(gotResult, input.wantResult) // eq = reflect.DeepEqual(gotResult, input.wantResult)
} // }
if !eq /*gotResult != input.wantResult*/ { if !eq /*gotResult != input.wantResult*/ {
t.Errorf(">>>%s/%d: `%s` -> result = %v [%s], want = %v [%s]", section, count, input.source, gotResult, kern.TypeName(gotResult), input.wantResult, kern.TypeName(input.wantResult)) t.Errorf(">>>%s/%d: `%s` -> result = %v [%s], want = %v [%s]", section, count, input.source, gotResult, kern.TypeName(gotResult), input.wantResult, kern.TypeName(input.wantResult))

View File

@ -38,8 +38,9 @@ func TestFractionsParser(t *testing.T) {
/* 23 */ {`1+1:2+0.5`, float64(2), nil}, /* 23 */ {`1+1:2+0.5`, float64(2), nil},
/* 24 */ {`1:(2-2)`, nil, `[1:3] division by zero`}, /* 24 */ {`1:(2-2)`, nil, `[1:3] division by zero`},
/* 25 */ {`[0,1][1-1]:1`, kern.NewFraction(0, 1), nil}, /* 25 */ {`[0,1][1-1]:1`, kern.NewFraction(0, 1), nil},
/* 26 */ {`1:2 == 0.5`, true, nil},
} }
// runTestSuiteSpec(t, section, inputs, 25) // runTestSuiteSpec(t, section, inputs, 26)
runTestSuite(t, section, inputs) runTestSuite(t, section, inputs)
} }

View File

@ -35,16 +35,84 @@ func TestOperator(t *testing.T) {
/* 20 */ {`a=1; a^=2`, int64(3), nil}, /* 20 */ {`a=1; a^=2`, int64(3), nil},
/* 21 */ {`a=1; ++a`, int64(2), nil}, /* 21 */ {`a=1; ++a`, int64(2), nil},
/* 22 */ {`a=1; --a`, int64(0), nil}, /* 22 */ {`a=1; --a`, int64(0), nil},
/* 23 */ {`[1,2,3] map var("_")`, kern.NewList([]any{int64(1), int64(2), int64(3)}), nil},
/* 24 */ {`[1,2,3] map $_`, kern.NewList([]any{int64(1), int64(2), int64(3)}), nil},
/* 25 */ {`[1,2,3,4] filter ($_ % 2 == 0)`, kern.NewList([]any{int64(2), int64(4)}), nil},
/* 26 */ {`max=0; [2,3,1] digest max=(($_ > max) ? {$_} :: {max})`, int64(3), nil},
/* 27 */ {`["a","b"] join ["x"]`, kern.NewList([]any{"a", "b", "x"}), nil},
/* 28 */ {`["a","b"] join ["x"-true]`, nil, `[1:21] left operand 'x' [string] and right operand 'true' [bool] are not compatible with operator "-"`},
} }
// t.Setenv("EXPR_PATH", ".") // t.Setenv("EXPR_PATH", ".")
// runTestSuiteSpec(t, section, inputs, 28) // runTestSuiteSpec(t, section, inputs, 22)
runTestSuite(t, section, inputs) runTestSuite(t, section, inputs)
} }
func TestOperatorMap(t *testing.T) {
section := "Operator-Map"
inputs := []inputType{
/* 1 */ {`a=1; --a`, int64(0), nil},
/* 2 */ {`[1,2,3] map var("_")`, kern.NewList([]any{int64(1), int64(2), int64(3)}), nil},
/* 3 */ {`[1,2,3] map $_`, kern.NewList([]any{int64(1), int64(2), int64(3)}), nil},
}
// runTestSuiteSpec(t, section, inputs, 3)
runTestSuite(t, section, inputs)
}
func TestOperatorFilter(t *testing.T) {
section := "Operator-Filter"
inputs := []inputType{
/* 1 */ {`[1,2,3,4] filter ($_ % 2 == 0)`, kern.NewList([]any{int64(2), int64(4)}), nil},
}
// runTestSuiteSpec(t, section, inputs, 1)
runTestSuite(t, section, inputs)
}
func TestOperatorDigest(t *testing.T) {
section := "Operator-Digest"
inputs := []inputType{
/* 1 */ {`max=0; [2,3,1] digest max=(($_ > max) ? {$_} :: {max})`, int64(3), nil},
}
// runTestSuiteSpec(t, section, inputs, 29)
runTestSuite(t, section, inputs)
}
func TestOperatorJoin(t *testing.T) {
section := "Operator-Join"
inputs := []inputType{
/* 1 */ {`["a","b"] join ["x"]`, kern.NewList([]any{"a", "b", "x"}), nil},
/* 2 */ {`["a","b"] join ["x"-true]`, nil, `[1:21] left operand 'x' [string] and right operand 'true' [bool] are not compatible with operator "-"`},
}
// runTestSuiteSpec(t, section, inputs, 2)
runTestSuite(t, section, inputs)
}
func TestOperatorGroupBy(t *testing.T) {
section := "Operator-GroupBy"
inputs := []inputType{
/* 1 */ {`L=[{"num": 1, "alpha": "one"}, {"num": 2, "alpha": "two"}, {"num": 3, "alpha": "three"}]; L groupby "num"`,
kern.NewDict(map[any]any{
"1": kern.NewListA(kern.NewDict(map[any]any{"num": int64(1), "alpha": "one"})),
"2": kern.NewListA(kern.NewDict(map[any]any{"num": int64(2), "alpha": "two"})),
"3": kern.NewListA(kern.NewDict(map[any]any{"num": int64(3), "alpha": "three"})),
}),
nil},
/* 2 */ {`cars = [{"model": "compas", "vendor": "jeep"}, {"model": "limited", "vendor": "jeep"}, {"model": "600", "vendor":"fiat"}]; cars groupby "vendor"`,
kern.NewDict(map[any]any{
"jeep": kern.NewListA(
kern.NewDict(map[any]any{"model": "compas", "vendor": "jeep"}),
kern.NewDict(map[any]any{"model": "limited", "vendor": "jeep"})),
"fiat": kern.NewListA(kern.NewDict(map[any]any{"model": "600", "vendor": "fiat"})),
}),
nil},
/* 3 */ {`[3,4,5] groupby $__`,
kern.NewDict(map[any]any{
"0": kern.NewListA(int64(3)),
"1": kern.NewListA(int64(4)),
"2": kern.NewListA(int64(5)),
}),
nil},
}
runTestSuiteSpec(t, section, inputs, 3)
// runTestSuite(t, section, inputs)
}

View File

@ -235,6 +235,10 @@ func (t *term) errDivisionByZero() error {
return t.tk.Errorf("division by zero") return t.tk.Errorf("division by zero")
} }
func (t *term) errKeyNotFound(key any) error {
return t.tk.Errorf("key '%v' not found", key)
}
func (t *term) Errorf(template string, args ...any) (err error) { func (t *term) Errorf(template string, args ...any) (err error) {
err = t.tk.Errorf(template, args...) err = t.tk.Errorf(template, args...)
return return