1
0
mirror of https://github.com/stedolan/jq.git synced 2024-05-11 05:55:39 +00:00

Remove the insane "fold" operation, replace with saner "reduce".

This commit is contained in:
Stephen Dolan
2013-05-16 15:07:53 +01:00
parent 8fbee891b3
commit bc42812715
7 changed files with 37 additions and 38 deletions

View File

@ -554,10 +554,10 @@ static const char* const jq_builtins[] = {
"def unique: group_by(.) | map(.[0]);",
"def max_by(f): _max_by_impl(map([f]));",
"def min_by(f): _min_by_impl(map([f]));",
"def add: fold null as $sum (.[] | $sum + .);",
"def add: reduce .[] as $x (null; . + $x);",
"def del(f): delpaths([path(f)]);",
"def _assign(paths; value): value as $v | fold . as $obj (path(paths) as $p | $obj | setpath($p; $v));",
"def _modify(paths; update): fold . as $obj (path(paths) as $p | $obj | setpath($p; getpath($p) | update));",
"def _assign(paths; value): value as $v | reduce path(paths) as $p (.; setpath($p; $v));",
"def _modify(paths; update): reduce path(paths) as $p (.; setpath($p; getpath($p) | update));",
"def recurse(f): ., (f | select(. != null) | recurse(f));",
"def to_entries: [keys[] as $k | {key: $k, value: .[$k]}];",
"def from_entries: map({(.key): .value}) | add;",

View File

@ -320,15 +320,24 @@ block gen_collect(block expr) {
gen_op_var_bound(LOADV, array_var));
}
block gen_fold(const char* varname, block init, block fold) {
block loop = BLOCK(fold, gen_op_var_unbound(STOREV, varname), gen_op_simple(BACKTRACK));
block gen_reduce(const char* varname, block source, block init, block body) {
block res_var = block_bind(gen_op_var_unbound(STOREV, "reduce"),
gen_noop(), OP_HAS_VARIABLE);
block loop = BLOCK(gen_op_simple(DUP),
source,
block_bind(gen_op_var_unbound(STOREV, varname),
BLOCK(gen_op_var_bound(LOADVN, res_var),
body,
gen_op_var_bound(STOREV, res_var)),
OP_HAS_VARIABLE),
gen_op_simple(BACKTRACK));
return BLOCK(gen_op_simple(DUP),
init,
block_bind(gen_op_var_unbound(STOREV, varname),
BLOCK(gen_op_target(FORK, loop),
loop,
gen_op_var_unbound(LOADV, varname)),
OP_HAS_VARIABLE));
res_var,
gen_op_target(FORK, loop),
loop,
gen_op_var_bound(LOADVN, res_var));
}
block gen_definedor(block a, block b) {

View File

@ -34,7 +34,7 @@ block gen_call(const char* name, block body);
block gen_subexp(block a);
block gen_both(block a, block b);
block gen_collect(block expr);
block gen_fold(const char* varname, block init, block body);
block gen_reduce(const char* varname, block source, block init, block body);
block gen_definedor(block a, block b);
block gen_condbranch(block iftrue, block iffalse);
block gen_and(block a, block b);

View File

@ -1016,7 +1016,7 @@ sections:
(many jq functions such as `map` and `find` are in fact written
in jq).
Finally, jq has a `fold` operation, which is very powerful but a
Finally, jq has a `reduce` operation, which is very powerful but a
bit tricky. Again, it's mostly used internally, to define some
useful bits of jq's standard library.
@ -1126,36 +1126,26 @@ sections:
input: '[[1,2],[10,20]]'
output: ['[[1,2,1,2], [10,20,1,2]]']
- title: Fold
- title: Reduce
body: |
The `fold` syntax in jq allows you to combine all of the
The `reduce` syntax in jq allows you to combine all of the
results of an expression by accumulating them into a single
answer. As an example, we'll pass `[3,2,1]` to this expression:
fold 0 as $sum (.[] | $sum + .)
reduce .[] as $item (0; . + $item)
The variable `$sum` is first given the value `0`. The body
of the fold (i.e. `.[] | $sum + .`) is evaluated. `.[]`
produces three results, `3`, `2`, and `1`. For the first
one, `$sum + .` gives `3`.
For each result that `.[]` produces, `. + $item` is run to
accumulate a running total, starting from 0. In this
example, `.[]` produces the results 3, 2, and 1, so the
effect is similar to running something like this:
Having produced this answer, jq backtracks to find the next
result as per usual. However, this time, `$sum` is set to
the previous value of the body, so `$sum + .` gives
`5`. After the final backtracking, `$sum + .` gives
`6`. This final value is used as the value of the entire
`fold` expression, so the above filter returns `6`.
More formally, in order to evaluate `fold INIT as $VAR
(BODY)`, jq first sets `$VAR` to the value of `INIT`. It
then runs through `BODY`. Each time `BODY` produces a value,
`$VAR` is set to that value and jq backtracks to find the
next one. When `BODY` stops producing values, the final
value of `$VAR` is the result of the entire expression.
0 | (3 as $item | . + $item) |
(2 as $item | . + $item) |
(1 as $item | . + $item)
examples:
- program: 'fold 0 as $sum (.[] | $sum + .)'
- program: 'reduce .[] as $item (0; . + $item)'
input: '[10,2,5,3]'
output: ['20']

View File

@ -49,7 +49,7 @@ struct lexer_param;
"and" { return AND; }
"or" { return OR; }
"end" { return END; }
"fold" { return FOLD; }
"reduce" { return REDUCE; }
"//" { return DEFINEDOR; }
"|=" { return SETPIPE; }
"+=" { return SETPLUS; }

View File

@ -58,7 +58,7 @@ struct lexer_param;
%token THEN "then"
%token ELSE "else"
%token ELSE_IF "elif"
%token FOLD "fold"
%token REDUCE "reduce"
%token END "end"
%token AND "and"
%token OR "or"
@ -226,8 +226,8 @@ Term "as" '$' IDENT '|' Exp {
jv_free($4);
} |
"fold" Term "as" '$' IDENT '(' Exp ')' {
$$ = gen_fold(jv_string_value($5), $2, $7);
"reduce" Term "as" '$' IDENT '(' Exp ';' Exp ')' {
$$ = gen_reduce(jv_string_value($5), $2, $7, $9);
jv_free($5);
} |

View File

@ -379,7 +379,7 @@ def fac: if . == 1 then 1 else . * (. - 1 | fac) end; [.[] | fac]
# []
# 1001
fold 0 as $s (.[] | $s + .)
reduce .[] as $x (0; . + $x)
[1,2,4]
7