rowpipe 1.2.0 → 2.1.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +238 -55
- package/dist/cli/commands/count.d.ts +12 -0
- package/dist/cli/commands/count.d.ts.map +1 -0
- package/dist/cli/commands/count.js +49 -0
- package/dist/cli/commands/count.js.map +1 -0
- package/dist/cli/commands/explain.d.ts +3 -0
- package/dist/cli/commands/explain.d.ts.map +1 -0
- package/dist/cli/commands/explain.js +11 -0
- package/dist/cli/commands/explain.js.map +1 -0
- package/dist/cli/commands/files.d.ts +3 -1
- package/dist/cli/commands/files.d.ts.map +1 -1
- package/dist/cli/commands/files.js +6 -2
- package/dist/cli/commands/files.js.map +1 -1
- package/dist/cli/commands/group.d.ts +12 -0
- package/dist/cli/commands/group.d.ts.map +1 -0
- package/dist/cli/commands/group.js +51 -0
- package/dist/cli/commands/group.js.map +1 -0
- package/dist/cli/commands/head.d.ts +7 -0
- package/dist/cli/commands/head.d.ts.map +1 -0
- package/dist/cli/commands/head.js +6 -0
- package/dist/cli/commands/head.js.map +1 -0
- package/dist/cli/commands/limit.d.ts +11 -0
- package/dist/cli/commands/limit.d.ts.map +1 -0
- package/dist/cli/commands/limit.js +40 -0
- package/dist/cli/commands/limit.js.map +1 -0
- package/dist/cli/commands/offset.d.ts +11 -0
- package/dist/cli/commands/offset.d.ts.map +1 -0
- package/dist/cli/commands/offset.js +39 -0
- package/dist/cli/commands/offset.js.map +1 -0
- package/dist/cli/commands/pipeline.d.ts +43 -0
- package/dist/cli/commands/pipeline.d.ts.map +1 -0
- package/dist/cli/commands/pipeline.js +198 -0
- package/dist/cli/commands/pipeline.js.map +1 -0
- package/dist/cli/commands/sort.d.ts +17 -0
- package/dist/cli/commands/sort.d.ts.map +1 -0
- package/dist/cli/commands/sort.js +55 -0
- package/dist/cli/commands/sort.js.map +1 -0
- package/dist/cli/commands/tail.d.ts +13 -0
- package/dist/cli/commands/tail.d.ts.map +1 -0
- package/dist/cli/commands/tail.js +39 -0
- package/dist/cli/commands/tail.js.map +1 -0
- package/dist/cli/commands/top.d.ts +19 -0
- package/dist/cli/commands/top.d.ts.map +1 -0
- package/dist/cli/commands/top.js +51 -0
- package/dist/cli/commands/top.js.map +1 -0
- package/dist/cli/commands/unique.d.ts +15 -0
- package/dist/cli/commands/unique.d.ts.map +1 -0
- package/dist/cli/commands/unique.js +44 -0
- package/dist/cli/commands/unique.js.map +1 -0
- package/dist/cli/index.js +253 -3
- package/dist/cli/index.js.map +1 -1
- package/dist/core/pipeline.d.ts +40 -0
- package/dist/core/pipeline.d.ts.map +1 -1
- package/dist/core/pipeline.js +65 -0
- package/dist/core/pipeline.js.map +1 -1
- package/dist/files/reader.d.ts.map +1 -1
- package/dist/files/reader.js +4 -0
- package/dist/files/reader.js.map +1 -1
- package/dist/files/types.d.ts +2 -0
- package/dist/files/types.d.ts.map +1 -1
- package/dist/index.d.ts +12 -0
- package/dist/index.d.ts.map +1 -1
- package/dist/index.js +13 -0
- package/dist/index.js.map +1 -1
- package/dist/planner/executor.d.ts +8 -0
- package/dist/planner/executor.d.ts.map +1 -0
- package/dist/planner/executor.js +86 -0
- package/dist/planner/executor.js.map +1 -0
- package/dist/planner/explain.d.ts +6 -0
- package/dist/planner/explain.d.ts.map +1 -0
- package/dist/planner/explain.js +91 -0
- package/dist/planner/explain.js.map +1 -0
- package/dist/planner/index.d.ts +5 -0
- package/dist/planner/index.d.ts.map +1 -0
- package/dist/planner/index.js +5 -0
- package/dist/planner/index.js.map +1 -0
- package/dist/planner/optimizer.d.ts +9 -0
- package/dist/planner/optimizer.d.ts.map +1 -0
- package/dist/planner/optimizer.js +78 -0
- package/dist/planner/optimizer.js.map +1 -0
- package/dist/planner/types.d.ts +63 -0
- package/dist/planner/types.d.ts.map +1 -0
- package/dist/planner/types.js +2 -0
- package/dist/planner/types.js.map +1 -0
- package/dist/transforms/count.d.ts +21 -0
- package/dist/transforms/count.d.ts.map +1 -0
- package/dist/transforms/count.js +65 -0
- package/dist/transforms/count.js.map +1 -0
- package/dist/transforms/expression.d.ts.map +1 -1
- package/dist/transforms/expression.js +15 -0
- package/dist/transforms/expression.js.map +1 -1
- package/dist/transforms/group.d.ts +25 -0
- package/dist/transforms/group.d.ts.map +1 -0
- package/dist/transforms/group.js +71 -0
- package/dist/transforms/group.js.map +1 -0
- package/dist/transforms/limit.d.ts +7 -0
- package/dist/transforms/limit.d.ts.map +1 -0
- package/dist/transforms/limit.js +40 -0
- package/dist/transforms/limit.js.map +1 -0
- package/dist/transforms/offset.d.ts +7 -0
- package/dist/transforms/offset.d.ts.map +1 -0
- package/dist/transforms/offset.js +38 -0
- package/dist/transforms/offset.js.map +1 -0
- package/dist/transforms/sort/comparator.d.ts +43 -0
- package/dist/transforms/sort/comparator.d.ts.map +1 -0
- package/dist/transforms/sort/comparator.js +131 -0
- package/dist/transforms/sort/comparator.js.map +1 -0
- package/dist/transforms/sort/external-sort.d.ts +8 -0
- package/dist/transforms/sort/external-sort.d.ts.map +1 -0
- package/dist/transforms/sort/external-sort.js +159 -0
- package/dist/transforms/sort/external-sort.js.map +1 -0
- package/dist/transforms/sort/index.d.ts +10 -0
- package/dist/transforms/sort/index.d.ts.map +1 -0
- package/dist/transforms/sort/index.js +11 -0
- package/dist/transforms/sort/index.js.map +1 -0
- package/dist/transforms/tail.d.ts +6 -0
- package/dist/transforms/tail.d.ts.map +1 -0
- package/dist/transforms/tail.js +27 -0
- package/dist/transforms/tail.js.map +1 -0
- package/dist/transforms/top.d.ts +17 -0
- package/dist/transforms/top.d.ts.map +1 -0
- package/dist/transforms/top.js +52 -0
- package/dist/transforms/top.js.map +1 -0
- package/dist/transforms/unique.d.ts +13 -0
- package/dist/transforms/unique.d.ts.map +1 -0
- package/dist/transforms/unique.js +91 -0
- package/dist/transforms/unique.js.map +1 -0
- package/dist/utils/formatting.d.ts +1 -0
- package/dist/utils/formatting.d.ts.map +1 -1
- package/dist/utils/formatting.js +34 -4
- package/dist/utils/formatting.js.map +1 -1
- package/dist/utils/heap.d.ts +25 -0
- package/dist/utils/heap.d.ts.map +1 -0
- package/dist/utils/heap.js +90 -0
- package/dist/utils/heap.js.map +1 -0
- package/dist/utils/keystore.d.ts +80 -0
- package/dist/utils/keystore.d.ts.map +1 -0
- package/dist/utils/keystore.js +217 -0
- package/dist/utils/keystore.js.map +1 -0
- package/dist/utils/ring-buffer.d.ts +15 -0
- package/dist/utils/ring-buffer.d.ts.map +1 -0
- package/dist/utils/ring-buffer.js +40 -0
- package/dist/utils/ring-buffer.js.map +1 -0
- package/dist/writers/markdown.d.ts.map +1 -1
- package/dist/writers/markdown.js +41 -12
- package/dist/writers/markdown.js.map +1 -1
- package/package.json +3 -2
- package/skills/rowpipe/SKILL.md +27 -9
|
@@ -0,0 +1,131 @@
|
|
|
1
|
+
/**
|
|
2
|
+
* Parses sort specifications from strings (e.g. "country,revenue:desc", "age:desc") or objects.
|
|
3
|
+
*/
|
|
4
|
+
export function parseSortSpecs(input, defaults = {}) {
|
|
5
|
+
const fallbackDir = defaults.defaultDirection || "asc";
|
|
6
|
+
if (Array.isArray(input) && input.length > 0 && typeof input[0] === "object") {
|
|
7
|
+
return input.map((spec) => ({
|
|
8
|
+
...spec,
|
|
9
|
+
direction: spec.direction || fallbackDir,
|
|
10
|
+
nulls: spec.nulls || defaults.nulls || "last",
|
|
11
|
+
ignoreCase: spec.ignoreCase ?? defaults.ignoreCase ?? false,
|
|
12
|
+
natural: spec.natural ?? defaults.natural ?? false,
|
|
13
|
+
}));
|
|
14
|
+
}
|
|
15
|
+
const rawSpecs = [];
|
|
16
|
+
if (typeof input === "string") {
|
|
17
|
+
rawSpecs.push(...input.split(",").map((s) => s.trim()).filter(Boolean));
|
|
18
|
+
}
|
|
19
|
+
else if (Array.isArray(input)) {
|
|
20
|
+
for (const item of input) {
|
|
21
|
+
if (typeof item === "string") {
|
|
22
|
+
rawSpecs.push(...item.split(",").map((s) => s.trim()).filter(Boolean));
|
|
23
|
+
}
|
|
24
|
+
}
|
|
25
|
+
}
|
|
26
|
+
return rawSpecs.map((spec) => {
|
|
27
|
+
const parts = spec.split(":");
|
|
28
|
+
const column = parts[0].trim();
|
|
29
|
+
let direction;
|
|
30
|
+
if (parts[1]) {
|
|
31
|
+
const dirStr = parts[1].trim().toLowerCase();
|
|
32
|
+
direction = dirStr === "desc" ? "desc" : "asc";
|
|
33
|
+
}
|
|
34
|
+
else {
|
|
35
|
+
direction = fallbackDir;
|
|
36
|
+
}
|
|
37
|
+
return {
|
|
38
|
+
column,
|
|
39
|
+
direction,
|
|
40
|
+
nulls: defaults.nulls || "last",
|
|
41
|
+
ignoreCase: defaults.ignoreCase ?? false,
|
|
42
|
+
natural: defaults.natural ?? false,
|
|
43
|
+
};
|
|
44
|
+
});
|
|
45
|
+
}
|
|
46
|
+
/**
|
|
47
|
+
* Compares two primitive values with typed semantics (numbers, dates, booleans, strings).
|
|
48
|
+
*/
|
|
49
|
+
export function compareValues(a, b, spec) {
|
|
50
|
+
const isANull = a === null || a === undefined || a === "";
|
|
51
|
+
const isBNull = b === null || b === undefined || b === "";
|
|
52
|
+
if (isANull && isBNull)
|
|
53
|
+
return 0;
|
|
54
|
+
if (isANull)
|
|
55
|
+
return spec.nulls === "first" ? -1 : 1;
|
|
56
|
+
if (isBNull)
|
|
57
|
+
return spec.nulls === "first" ? 1 : -1;
|
|
58
|
+
// Number comparison (including numeric strings)
|
|
59
|
+
const isANum = typeof a === "number" || (typeof a === "string" && /^-?\d+(\.\d+)?$/.test(a.trim()));
|
|
60
|
+
const isBNum = typeof b === "number" || (typeof b === "string" && /^-?\d+(\.\d+)?$/.test(b.trim()));
|
|
61
|
+
if (isANum && isBNum) {
|
|
62
|
+
const numA = typeof a === "number" ? a : Number.parseFloat(a);
|
|
63
|
+
const numB = typeof b === "number" ? b : Number.parseFloat(b);
|
|
64
|
+
if (!Number.isNaN(numA) && !Number.isNaN(numB)) {
|
|
65
|
+
return numA < numB ? -1 : numA > numB ? 1 : 0;
|
|
66
|
+
}
|
|
67
|
+
}
|
|
68
|
+
// Boolean comparison
|
|
69
|
+
if (typeof a === "boolean" || typeof b === "boolean") {
|
|
70
|
+
const boolA = Boolean(a);
|
|
71
|
+
const boolB = Boolean(b);
|
|
72
|
+
return boolA === boolB ? 0 : boolA ? 1 : -1;
|
|
73
|
+
}
|
|
74
|
+
// Date comparison
|
|
75
|
+
if (a instanceof Date && b instanceof Date) {
|
|
76
|
+
return a.getTime() - b.getTime();
|
|
77
|
+
}
|
|
78
|
+
// String comparison
|
|
79
|
+
const strA = String(a);
|
|
80
|
+
const strB = String(b);
|
|
81
|
+
if (spec.natural) {
|
|
82
|
+
return strA.localeCompare(strB, undefined, {
|
|
83
|
+
numeric: true,
|
|
84
|
+
sensitivity: spec.ignoreCase ? "base" : "variant",
|
|
85
|
+
});
|
|
86
|
+
}
|
|
87
|
+
if (spec.ignoreCase) {
|
|
88
|
+
const lowerA = strA.toLowerCase();
|
|
89
|
+
const lowerB = strB.toLowerCase();
|
|
90
|
+
return lowerA < lowerB ? -1 : lowerA > lowerB ? 1 : 0;
|
|
91
|
+
}
|
|
92
|
+
return strA < strB ? -1 : strA > strB ? 1 : 0;
|
|
93
|
+
}
|
|
94
|
+
/**
|
|
95
|
+
* Creates a deterministic, multi-column, stable row comparator.
|
|
96
|
+
*/
|
|
97
|
+
export function createRowComparator(specs) {
|
|
98
|
+
return (aObj, bObj) => {
|
|
99
|
+
const a = aObj.row;
|
|
100
|
+
const b = bObj.row;
|
|
101
|
+
for (let i = 0; i < specs.length; i++) {
|
|
102
|
+
const spec = specs[i];
|
|
103
|
+
const valA = a[spec.column];
|
|
104
|
+
const valB = b[spec.column];
|
|
105
|
+
const cmp = compareValues(valA, valB, spec);
|
|
106
|
+
if (cmp !== 0) {
|
|
107
|
+
return spec.direction === "desc" ? -cmp : cmp;
|
|
108
|
+
}
|
|
109
|
+
}
|
|
110
|
+
// Preserve original insertion sequence for stability
|
|
111
|
+
return aObj.seq - bObj.seq;
|
|
112
|
+
};
|
|
113
|
+
}
|
|
114
|
+
/**
|
|
115
|
+
* Creates a raw Row comparator (without seq).
|
|
116
|
+
*/
|
|
117
|
+
export function createRawRowComparator(specs) {
|
|
118
|
+
return (a, b) => {
|
|
119
|
+
for (let i = 0; i < specs.length; i++) {
|
|
120
|
+
const spec = specs[i];
|
|
121
|
+
const valA = a[spec.column];
|
|
122
|
+
const valB = b[spec.column];
|
|
123
|
+
const cmp = compareValues(valA, valB, spec);
|
|
124
|
+
if (cmp !== 0) {
|
|
125
|
+
return spec.direction === "desc" ? -cmp : cmp;
|
|
126
|
+
}
|
|
127
|
+
}
|
|
128
|
+
return 0;
|
|
129
|
+
};
|
|
130
|
+
}
|
|
131
|
+
//# sourceMappingURL=comparator.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"comparator.js","sourceRoot":"","sources":["../../../src/transforms/sort/comparator.ts"],"names":[],"mappings":"AAyBA;;GAEG;AACH,MAAM,UAAU,cAAc,CAC5B,KAAwC,EACxC,WAKI,EAAE;IAEN,MAAM,WAAW,GAAG,QAAQ,CAAC,gBAAgB,IAAI,KAAK,CAAC;IACvD,IAAI,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,QAAQ,EAAE,CAAC;QAC7E,OAAQ,KAAuB,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;YAC7C,GAAG,IAAI;YACP,SAAS,EAAE,IAAI,CAAC,SAAS,IAAI,WAAW;YACxC,KAAK,EAAE,IAAI,CAAC,KAAK,IAAI,QAAQ,CAAC,KAAK,IAAI,MAAM;YAC7C,UAAU,EAAE,IAAI,CAAC,UAAU,IAAI,QAAQ,CAAC,UAAU,IAAI,KAAK;YAC3D,OAAO,EAAE,IAAI,CAAC,OAAO,IAAI,QAAQ,CAAC,OAAO,IAAI,KAAK;SACnD,CAAC,CAAC,CAAC;IACN,CAAC;IAED,MAAM,QAAQ,GAAa,EAAE,CAAC;IAC9B,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;QAC9B,QAAQ,CAAC,IAAI,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC;IAC1E,CAAC;SAAM,IAAI,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC;QAChC,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;YACzB,IAAI,OAAO,IAAI,KAAK,QAAQ,EAAE,CAAC;gBAC7B,QAAQ,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC;YACzE,CAAC;QACH,CAAC;IACH,CAAC;IAED,OAAO,QAAQ,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE;QAC3B,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;QAC9B,MAAM,MAAM,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;QAChC,IAAI,SAAyB,CAAC;QAC9B,IAAI,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC;YACb,MAAM,MAAM,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;YAC7C,SAAS,GAAG,MAAM,KAAK,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC;QACjD,CAAC;aAAM,CAAC;YACN,SAAS,GAAG,WAAW,CAAC;QAC1B,CAAC;QAED,OAAO;YACL,MAAM;YACN,SAAS;YACT,KAAK,EAAE,QAAQ,CAAC,KAAK,IAAI,MAAM;YAC/B,UAAU,EAAE,QAAQ,CAAC,UAAU,IAAI,KAAK;YACxC,OAAO,EAAE,QAAQ,CAAC,OAAO,IAAI,KAAK;SACnC,CAAC;IACJ,CAAC,CAAC,CAAC;AACL,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,aAAa,CAC3B,CAAU,EACV,CAAU,EACV,IAAiB;IAEjB,MAAM,OAAO,GAAG,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,SAAS,IAAI,CAAC,KAAK,EAAE,CAAC;IAC1D,MAAM,OAAO,GAAG,CAAC,KAAK,IAAI,IAAI,CAAC,KAAK,SAAS,IAAI,CAAC,KAAK,EAAE,CAAC;IAE1D,IAAI,OAAO,IAAI,OAAO;QAAE,OAAO,CAAC,CAAC;IACjC,IAAI,OAAO;QAAE,OAAO,IAAI,CAAC,KAAK,KAAK,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACpD,IAAI,OAAO;QAAE,OAAO,IAAI,CAAC,KAAK,KAAK,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAEpD,gDAAgD;IAChD,MAAM,MAAM,GAAG,OAAO,CAAC,KAAK,QAAQ,IAAI,CAAC,OAAO,CAAC,KAAK,QAAQ,IAAI,iBAAiB,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;IACpG,MAAM,MAAM,GAAG,OAAO,CAAC,KAAK,QAAQ,IAAI,CAAC,OAAO,CAAC,KAAK,QAAQ,IAAI,iBAAiB,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;IAEpG,IAAI,MAAM,IAAI,MAAM,EAAE,CAAC;QACrB,MAAM,IAAI,GAAG,OAAO,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,UAAU,CAAC,CAAW,CAAC,CAAC;QACxE,MAAM,IAAI,GAAG,OAAO,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,UAAU,CAAC,CAAW,CAAC,CAAC;QACxE,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC;YAC/C,OAAO,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QAChD,CAAC;IACH,CAAC;IAED,qBAAqB;IACrB,IAAI,OAAO,CAAC,KAAK,SAAS,IAAI,OAAO,CAAC,KAAK,SAAS,EAAE,CAAC;QACrD,MAAM,KAAK,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;QACzB,MAAM,KAAK,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC;QACzB,OAAO,KAAK,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAC9C,CAAC;IAED,kBAAkB;IAClB,IAAI,CAAC,YAAY,IAAI,IAAI,CAAC,YAAY,IAAI,EAAE,CAAC;QAC3C,OAAO,CAAC,CAAC,OAAO,EAAE,GAAG,CAAC,CAAC,OAAO,EAAE,CAAC;IACnC,CAAC;IAED,oBAAoB;IACpB,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;IACvB,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC;IAEvB,IAAI,IAAI,CAAC,OAAO,EAAE,CAAC;QACjB,OAAO,IAAI,CAAC,aAAa,CAAC,IAAI,EAAE,SAAS,EAAE;YACzC,OAAO,EAAE,IAAI;YACb,WAAW,EAAE,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,SAAS;SAClD,CAAC,CAAC;IACL,CAAC;IAED,IAAI,IAAI,CAAC,UAAU,EAAE,CAAC;QACpB,MAAM,MAAM,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC;QAClC,MAAM,MAAM,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC;QAClC,OAAO,MAAM,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,MAAM,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACxD,CAAC;IAED,OAAO,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AAChD,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,mBAAmB,CAAC,KAAoB;IACtD,OAAO,CAAC,IAAkB,EAAE,IAAkB,EAAU,EAAE;QACxD,MAAM,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC;QACnB,MAAM,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC;QAEnB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACtC,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC;YACvB,MAAM,IAAI,GAAG,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YAC5B,MAAM,IAAI,GAAG,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YAE5B,MAAM,GAAG,GAAG,aAAa,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC;YAC5C,IAAI,GAAG,KAAK,CAAC,EAAE,CAAC;gBACd,OAAO,IAAI,CAAC,SAAS,KAAK,MAAM,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC;YAChD,CAAC;QACH,CAAC;QAED,qDAAqD;QACrD,OAAO,IAAI,CAAC,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC;IAC7B,CAAC,CAAC;AACJ,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,sBAAsB,CAAC,KAAoB;IACzD,OAAO,CAAC,CAAM,EAAE,CAAM,EAAU,EAAE;QAChC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;YACtC,MAAM,IAAI,GAAG,KAAK,CAAC,CAAC,CAAE,CAAC;YACvB,MAAM,IAAI,GAAG,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YAC5B,MAAM,IAAI,GAAG,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;YAE5B,MAAM,GAAG,GAAG,aAAa,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC;YAC5C,IAAI,GAAG,KAAK,CAAC,EAAE,CAAC;gBACd,OAAO,IAAI,CAAC,SAAS,KAAK,MAAM,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC;YAChD,CAAC;QACH,CAAC;QACD,OAAO,CAAC,CAAC;IACX,CAAC,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1,8 @@
|
|
|
1
|
+
import type { TransformFunction } from "../../core/types.js";
|
|
2
|
+
import { SortOptions } from "./comparator.js";
|
|
3
|
+
/**
|
|
4
|
+
* External Merge Sort Engine.
|
|
5
|
+
* Sorts arbitrary-scale datasets with bounded RAM and K-Way Heap merge.
|
|
6
|
+
*/
|
|
7
|
+
export declare function externalSort(options: SortOptions): TransformFunction;
|
|
8
|
+
//# sourceMappingURL=external-sort.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"external-sort.d.ts","sourceRoot":"","sources":["../../../src/transforms/sort/external-sort.ts"],"names":[],"mappings":"AAKA,OAAO,KAAK,EAA8B,iBAAiB,EAAE,MAAM,qBAAqB,CAAC;AAEzF,OAAO,EAIL,WAAW,EACZ,MAAM,iBAAiB,CAAC;AAQzB;;;GAGG;AACH,wBAAgB,YAAY,CAAC,OAAO,EAAE,WAAW,GAAG,iBAAiB,CAwKpE"}
|
|
@@ -0,0 +1,159 @@
|
|
|
1
|
+
import { createReadStream, createWriteStream, promises as fsPromises } from "node:fs";
|
|
2
|
+
import * as os from "node:os";
|
|
3
|
+
import * as path from "node:path";
|
|
4
|
+
import * as readline from "node:readline";
|
|
5
|
+
import { parseMemoryLimit } from "../../diff/storage/spillable-index.js";
|
|
6
|
+
import { Heap } from "../../utils/heap.js";
|
|
7
|
+
import { createRowComparator, parseSortSpecs, } from "./comparator.js";
|
|
8
|
+
/**
|
|
9
|
+
* External Merge Sort Engine.
|
|
10
|
+
* Sorts arbitrary-scale datasets with bounded RAM and K-Way Heap merge.
|
|
11
|
+
*/
|
|
12
|
+
export function externalSort(options) {
|
|
13
|
+
const specs = parseSortSpecs(options.by, {
|
|
14
|
+
nulls: options.nulls,
|
|
15
|
+
ignoreCase: options.ignoreCase,
|
|
16
|
+
natural: options.natural,
|
|
17
|
+
});
|
|
18
|
+
const rowComparator = createRowComparator(specs);
|
|
19
|
+
const memoryLimitBytes = parseMemoryLimit(options.memoryLimit, 256 * 1024 * 1024);
|
|
20
|
+
const effectiveBatchSize = options.batchSize || 1000;
|
|
21
|
+
const baseTempDir = options.tempDir || os.tmpdir();
|
|
22
|
+
return (stream) => {
|
|
23
|
+
return (async function* () {
|
|
24
|
+
let currentRun = [];
|
|
25
|
+
let currentRunBytes = 0;
|
|
26
|
+
let totalSeq = 0;
|
|
27
|
+
const runFiles = [];
|
|
28
|
+
const sessionDir = path.join(baseTempDir, `rowpipe-sort-${Date.now()}-${Math.random().toString(36).slice(2, 8)}`);
|
|
29
|
+
const cleanup = async () => {
|
|
30
|
+
try {
|
|
31
|
+
await fsPromises.rm(sessionDir, { recursive: true, force: true });
|
|
32
|
+
}
|
|
33
|
+
catch {
|
|
34
|
+
// Ignore temp cleanup error
|
|
35
|
+
}
|
|
36
|
+
};
|
|
37
|
+
try {
|
|
38
|
+
for await (const batch of stream) {
|
|
39
|
+
for (let i = 0; i < batch.rows.length; i++) {
|
|
40
|
+
const row = batch.rows[i];
|
|
41
|
+
currentRun.push({ row, seq: totalSeq++ });
|
|
42
|
+
// Estimate row memory (~64 bytes overhead + keys/values)
|
|
43
|
+
currentRunBytes += 64 + Object.keys(row).length * 24;
|
|
44
|
+
if (currentRunBytes >= memoryLimitBytes) {
|
|
45
|
+
// Spill current run
|
|
46
|
+
await fsPromises.mkdir(sessionDir, { recursive: true });
|
|
47
|
+
currentRun.sort(rowComparator);
|
|
48
|
+
const runFilePath = path.join(sessionDir, `run_${runFiles.length}.jsonl`);
|
|
49
|
+
const writeStream = createWriteStream(runFilePath, { encoding: "utf-8" });
|
|
50
|
+
for (let r = 0; r < currentRun.length; r++) {
|
|
51
|
+
const item = currentRun[r];
|
|
52
|
+
writeStream.write(JSON.stringify({ row: item.row, seq: item.seq }) + "\n");
|
|
53
|
+
}
|
|
54
|
+
await new Promise((resolve, reject) => {
|
|
55
|
+
writeStream.end((err) => {
|
|
56
|
+
if (err)
|
|
57
|
+
reject(err);
|
|
58
|
+
else
|
|
59
|
+
resolve();
|
|
60
|
+
});
|
|
61
|
+
});
|
|
62
|
+
runFiles.push(runFilePath);
|
|
63
|
+
currentRun = [];
|
|
64
|
+
currentRunBytes = 0;
|
|
65
|
+
}
|
|
66
|
+
}
|
|
67
|
+
}
|
|
68
|
+
// Case 1: All data fit in memory without spilling
|
|
69
|
+
if (runFiles.length === 0) {
|
|
70
|
+
currentRun.sort(rowComparator);
|
|
71
|
+
let offset = 0;
|
|
72
|
+
for (let i = 0; i < currentRun.length; i += effectiveBatchSize) {
|
|
73
|
+
const slice = currentRun.slice(i, i + effectiveBatchSize).map((item) => item.row);
|
|
74
|
+
yield {
|
|
75
|
+
rows: slice,
|
|
76
|
+
offset,
|
|
77
|
+
};
|
|
78
|
+
offset += slice.length;
|
|
79
|
+
}
|
|
80
|
+
return;
|
|
81
|
+
}
|
|
82
|
+
// Case 2: Spill remainder run if any rows are in memory
|
|
83
|
+
if (currentRun.length > 0) {
|
|
84
|
+
await fsPromises.mkdir(sessionDir, { recursive: true });
|
|
85
|
+
currentRun.sort(rowComparator);
|
|
86
|
+
const runFilePath = path.join(sessionDir, `run_${runFiles.length}.jsonl`);
|
|
87
|
+
const writeStream = createWriteStream(runFilePath, { encoding: "utf-8" });
|
|
88
|
+
for (let r = 0; r < currentRun.length; r++) {
|
|
89
|
+
const item = currentRun[r];
|
|
90
|
+
writeStream.write(JSON.stringify({ row: item.row, seq: item.seq }) + "\n");
|
|
91
|
+
}
|
|
92
|
+
await new Promise((resolve, reject) => {
|
|
93
|
+
writeStream.end((err) => {
|
|
94
|
+
if (err)
|
|
95
|
+
reject(err);
|
|
96
|
+
else
|
|
97
|
+
resolve();
|
|
98
|
+
});
|
|
99
|
+
});
|
|
100
|
+
runFiles.push(runFilePath);
|
|
101
|
+
currentRun = [];
|
|
102
|
+
currentRunBytes = 0;
|
|
103
|
+
}
|
|
104
|
+
// Case 3: K-Way Merge using Min-Heap
|
|
105
|
+
const k = runFiles.length;
|
|
106
|
+
const iterators = new Array(k);
|
|
107
|
+
for (let i = 0; i < k; i++) {
|
|
108
|
+
const rl = readline.createInterface({
|
|
109
|
+
input: createReadStream(runFiles[i], { encoding: "utf-8" }),
|
|
110
|
+
crlfDelay: Infinity,
|
|
111
|
+
});
|
|
112
|
+
iterators[i] = rl[Symbol.asyncIterator]();
|
|
113
|
+
}
|
|
114
|
+
const heapComparator = (a, b) => {
|
|
115
|
+
return rowComparator({ row: a.row, seq: a.seq }, { row: b.row, seq: b.seq });
|
|
116
|
+
};
|
|
117
|
+
const heap = new Heap(heapComparator);
|
|
118
|
+
// Prime the heap with 1 item from each run
|
|
119
|
+
for (let i = 0; i < k; i++) {
|
|
120
|
+
const next = await iterators[i].next();
|
|
121
|
+
if (!next.done && next.value.trim().length > 0) {
|
|
122
|
+
const parsed = JSON.parse(next.value);
|
|
123
|
+
heap.push({ row: parsed.row, seq: parsed.seq, runIndex: i });
|
|
124
|
+
}
|
|
125
|
+
}
|
|
126
|
+
let outputRows = [];
|
|
127
|
+
let globalOffset = 0;
|
|
128
|
+
while (!heap.isEmpty()) {
|
|
129
|
+
const smallest = heap.pop();
|
|
130
|
+
outputRows.push(smallest.row);
|
|
131
|
+
if (outputRows.length >= effectiveBatchSize) {
|
|
132
|
+
yield {
|
|
133
|
+
rows: outputRows,
|
|
134
|
+
offset: globalOffset,
|
|
135
|
+
};
|
|
136
|
+
globalOffset += outputRows.length;
|
|
137
|
+
outputRows = [];
|
|
138
|
+
}
|
|
139
|
+
// Advance iterator for smallest.runIndex
|
|
140
|
+
const next = await iterators[smallest.runIndex].next();
|
|
141
|
+
if (!next.done && next.value.trim().length > 0) {
|
|
142
|
+
const parsed = JSON.parse(next.value);
|
|
143
|
+
heap.push({ row: parsed.row, seq: parsed.seq, runIndex: smallest.runIndex });
|
|
144
|
+
}
|
|
145
|
+
}
|
|
146
|
+
if (outputRows.length > 0) {
|
|
147
|
+
yield {
|
|
148
|
+
rows: outputRows,
|
|
149
|
+
offset: globalOffset,
|
|
150
|
+
};
|
|
151
|
+
}
|
|
152
|
+
}
|
|
153
|
+
finally {
|
|
154
|
+
await cleanup();
|
|
155
|
+
}
|
|
156
|
+
})();
|
|
157
|
+
};
|
|
158
|
+
}
|
|
159
|
+
//# sourceMappingURL=external-sort.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"external-sort.js","sourceRoot":"","sources":["../../../src/transforms/sort/external-sort.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,gBAAgB,EAAE,iBAAiB,EAAE,QAAQ,IAAI,UAAU,EAAE,MAAM,SAAS,CAAC;AACtF,OAAO,KAAK,EAAE,MAAM,SAAS,CAAC;AAC9B,OAAO,KAAK,IAAI,MAAM,WAAW,CAAC;AAClC,OAAO,KAAK,QAAQ,MAAM,eAAe,CAAC;AAC1C,OAAO,EAAE,gBAAgB,EAAE,MAAM,uCAAuC,CAAC;AAEzE,OAAO,EAAE,IAAI,EAAE,MAAM,qBAAqB,CAAC;AAC3C,OAAO,EACL,mBAAmB,EACnB,cAAc,GAGf,MAAM,iBAAiB,CAAC;AAQzB;;;GAGG;AACH,MAAM,UAAU,YAAY,CAAC,OAAoB;IAC/C,MAAM,KAAK,GAAG,cAAc,CAAC,OAAO,CAAC,EAAE,EAAE;QACvC,KAAK,EAAE,OAAO,CAAC,KAAK;QACpB,UAAU,EAAE,OAAO,CAAC,UAAU;QAC9B,OAAO,EAAE,OAAO,CAAC,OAAO;KACzB,CAAC,CAAC;IAEH,MAAM,aAAa,GAAG,mBAAmB,CAAC,KAAK,CAAC,CAAC;IACjD,MAAM,gBAAgB,GAAG,gBAAgB,CAAC,OAAO,CAAC,WAAW,EAAE,GAAG,GAAG,IAAI,GAAG,IAAI,CAAC,CAAC;IAClF,MAAM,kBAAkB,GAAG,OAAO,CAAC,SAAS,IAAI,IAAI,CAAC;IACrD,MAAM,WAAW,GAAG,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,MAAM,EAAE,CAAC;IAEnD,OAAO,CAAC,MAAkB,EAAc,EAAE;QACxC,OAAO,CAAC,KAAK,SAAS,CAAC;YACrB,IAAI,UAAU,GAAmB,EAAE,CAAC;YACpC,IAAI,eAAe,GAAG,CAAC,CAAC;YACxB,IAAI,QAAQ,GAAG,CAAC,CAAC;YACjB,MAAM,QAAQ,GAAa,EAAE,CAAC;YAC9B,MAAM,UAAU,GAAG,IAAI,CAAC,IAAI,CAC1B,WAAW,EACX,gBAAgB,IAAI,CAAC,GAAG,EAAE,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,CACvE,CAAC;YAEF,MAAM,OAAO,GAAG,KAAK,IAAI,EAAE;gBACzB,IAAI,CAAC;oBACH,MAAM,UAAU,CAAC,EAAE,CAAC,UAAU,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,KAAK,EAAE,IAAI,EAAE,CAAC,CAAC;gBACpE,CAAC;gBAAC,MAAM,CAAC;oBACP,4BAA4B;gBAC9B,CAAC;YACH,CAAC,CAAC;YAEF,IAAI,CAAC;gBACH,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;oBACjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;wBAC3C,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC;wBAC3B,UAAU,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,GAAG,EAAE,QAAQ,EAAE,EAAE,CAAC,CAAC;wBAC1C,yDAAyD;wBACzD,eAAe,IAAI,EAAE,GAAG,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,MAAM,GAAG,EAAE,CAAC;wBAErD,IAAI,eAAe,IAAI,gBAAgB,EAAE,CAAC;4BACxC,oBAAoB;4BACpB,MAAM,UAAU,CAAC,KAAK,CAAC,UAAU,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;4BACxD,UAAU,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;4BAE/B,MAAM,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,OAAO,QAAQ,CAAC,MAAM,QAAQ,CAAC,CAAC;4BAC1E,MAAM,WAAW,GAAG,iBAAiB,CAAC,WAAW,EAAE,EAAE,QAAQ,EAAE,OAAO,EAAE,CAAC,CAAC;4BAE1E,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gCAC3C,MAAM,IAAI,GAAG,UAAU,CAAC,CAAC,CAAE,CAAC;gCAC5B,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE,CAAC,GAAG,IAAI,CAAC,CAAC;4BAC7E,CAAC;4BAED,MAAM,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;gCAC1C,WAAW,CAAC,GAAG,CAAC,CAAC,GAAkB,EAAE,EAAE;oCACrC,IAAI,GAAG;wCAAE,MAAM,CAAC,GAAG,CAAC,CAAC;;wCAChB,OAAO,EAAE,CAAC;gCACjB,CAAC,CAAC,CAAC;4BACL,CAAC,CAAC,CAAC;4BAEH,QAAQ,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC;4BAC3B,UAAU,GAAG,EAAE,CAAC;4BAChB,eAAe,GAAG,CAAC,CAAC;wBACtB,CAAC;oBACH,CAAC;gBACH,CAAC;gBAED,kDAAkD;gBAClD,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;oBAC1B,UAAU,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;oBAC/B,IAAI,MAAM,GAAG,CAAC,CAAC;oBACf,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC,IAAI,kBAAkB,EAAE,CAAC;wBAC/D,MAAM,KAAK,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,GAAG,kBAAkB,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;wBAClF,MAAM;4BACJ,IAAI,EAAE,KAAK;4BACX,MAAM;yBACP,CAAC;wBACF,MAAM,IAAI,KAAK,CAAC,MAAM,CAAC;oBACzB,CAAC;oBACD,OAAO;gBACT,CAAC;gBAED,wDAAwD;gBACxD,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAC1B,MAAM,UAAU,CAAC,KAAK,CAAC,UAAU,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;oBACxD,UAAU,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;oBAE/B,MAAM,WAAW,GAAG,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,OAAO,QAAQ,CAAC,MAAM,QAAQ,CAAC,CAAC;oBAC1E,MAAM,WAAW,GAAG,iBAAiB,CAAC,WAAW,EAAE,EAAE,QAAQ,EAAE,OAAO,EAAE,CAAC,CAAC;oBAE1E,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;wBAC3C,MAAM,IAAI,GAAG,UAAU,CAAC,CAAC,CAAE,CAAC;wBAC5B,WAAW,CAAC,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,EAAE,IAAI,CAAC,GAAG,EAAE,CAAC,GAAG,IAAI,CAAC,CAAC;oBAC7E,CAAC;oBAED,MAAM,IAAI,OAAO,CAAO,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;wBAC1C,WAAW,CAAC,GAAG,CAAC,CAAC,GAAkB,EAAE,EAAE;4BACrC,IAAI,GAAG;gCAAE,MAAM,CAAC,GAAG,CAAC,CAAC;;gCAChB,OAAO,EAAE,CAAC;wBACjB,CAAC,CAAC,CAAC;oBACL,CAAC,CAAC,CAAC;oBAEH,QAAQ,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC;oBAC3B,UAAU,GAAG,EAAE,CAAC;oBAChB,eAAe,GAAG,CAAC,CAAC;gBACtB,CAAC;gBAED,qCAAqC;gBACrC,MAAM,CAAC,GAAG,QAAQ,CAAC,MAAM,CAAC;gBAC1B,MAAM,SAAS,GAA4B,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC;gBAExD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;oBAC3B,MAAM,EAAE,GAAG,QAAQ,CAAC,eAAe,CAAC;wBAClC,KAAK,EAAE,gBAAgB,CAAC,QAAQ,CAAC,CAAC,CAAE,EAAE,EAAE,QAAQ,EAAE,OAAO,EAAE,CAAC;wBAC5D,SAAS,EAAE,QAAQ;qBACpB,CAAC,CAAC;oBACH,SAAS,CAAC,CAAC,CAAC,GAAG,EAAE,CAAC,MAAM,CAAC,aAAa,CAAC,EAAE,CAAC;gBAC5C,CAAC;gBAED,MAAM,cAAc,GAAG,CAAC,CAAe,EAAE,CAAe,EAAE,EAAE;oBAC1D,OAAO,aAAa,CAAC,EAAE,GAAG,EAAE,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,CAAC,CAAC,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,CAAC,CAAC,GAAG,EAAE,CAAC,CAAC;gBAC/E,CAAC,CAAC;gBAEF,MAAM,IAAI,GAAG,IAAI,IAAI,CAAe,cAAc,CAAC,CAAC;gBAEpD,2CAA2C;gBAC3C,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;oBAC3B,MAAM,IAAI,GAAG,MAAM,SAAS,CAAC,CAAC,CAAE,CAAC,IAAI,EAAE,CAAC;oBACxC,IAAI,CAAC,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;wBAC/C,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAA8B,CAAC;wBACnE,IAAI,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,MAAM,CAAC,GAAG,EAAE,GAAG,EAAE,MAAM,CAAC,GAAG,EAAE,QAAQ,EAAE,CAAC,EAAE,CAAC,CAAC;oBAC/D,CAAC;gBACH,CAAC;gBAED,IAAI,UAAU,GAAU,EAAE,CAAC;gBAC3B,IAAI,YAAY,GAAG,CAAC,CAAC;gBAErB,OAAO,CAAC,IAAI,CAAC,OAAO,EAAE,EAAE,CAAC;oBACvB,MAAM,QAAQ,GAAG,IAAI,CAAC,GAAG,EAAG,CAAC;oBAC7B,UAAU,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC;oBAE9B,IAAI,UAAU,CAAC,MAAM,IAAI,kBAAkB,EAAE,CAAC;wBAC5C,MAAM;4BACJ,IAAI,EAAE,UAAU;4BAChB,MAAM,EAAE,YAAY;yBACrB,CAAC;wBACF,YAAY,IAAI,UAAU,CAAC,MAAM,CAAC;wBAClC,UAAU,GAAG,EAAE,CAAC;oBAClB,CAAC;oBAED,yCAAyC;oBACzC,MAAM,IAAI,GAAG,MAAM,SAAS,CAAC,QAAQ,CAAC,QAAQ,CAAE,CAAC,IAAI,EAAE,CAAC;oBACxD,IAAI,CAAC,IAAI,CAAC,IAAI,IAAI,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;wBAC/C,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAA8B,CAAC;wBACnE,IAAI,CAAC,IAAI,CAAC,EAAE,GAAG,EAAE,MAAM,CAAC,GAAG,EAAE,GAAG,EAAE,MAAM,CAAC,GAAG,EAAE,QAAQ,EAAE,QAAQ,CAAC,QAAQ,EAAE,CAAC,CAAC;oBAC/E,CAAC;gBACH,CAAC;gBAED,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;oBAC1B,MAAM;wBACJ,IAAI,EAAE,UAAU;wBAChB,MAAM,EAAE,YAAY;qBACrB,CAAC;gBACJ,CAAC;YACH,CAAC;oBAAS,CAAC;gBACT,MAAM,OAAO,EAAE,CAAC;YAClB,CAAC;QACH,CAAC,CAAC,EAAE,CAAC;IACP,CAAC,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1,10 @@
|
|
|
1
|
+
import type { TransformFunction } from "../../core/types.js";
|
|
2
|
+
import type { SortOptions } from "./comparator.js";
|
|
3
|
+
export * from "./comparator.js";
|
|
4
|
+
export * from "./external-sort.js";
|
|
5
|
+
/**
|
|
6
|
+
* Sorts tabular stream by one or more columns with typed comparisons,
|
|
7
|
+
* stable ordering, and automatic external merge sort on large streams.
|
|
8
|
+
*/
|
|
9
|
+
export declare function sortRows(options: SortOptions): TransformFunction;
|
|
10
|
+
//# sourceMappingURL=index.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/transforms/sort/index.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,iBAAiB,EAAE,MAAM,qBAAqB,CAAC;AAE7D,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAC;AAEnD,cAAc,iBAAiB,CAAC;AAChC,cAAc,oBAAoB,CAAC;AAEnC;;;GAGG;AACH,wBAAgB,QAAQ,CAAC,OAAO,EAAE,WAAW,GAAG,iBAAiB,CAEhE"}
|
|
@@ -0,0 +1,11 @@
|
|
|
1
|
+
import { externalSort } from "./external-sort.js";
|
|
2
|
+
export * from "./comparator.js";
|
|
3
|
+
export * from "./external-sort.js";
|
|
4
|
+
/**
|
|
5
|
+
* Sorts tabular stream by one or more columns with typed comparisons,
|
|
6
|
+
* stable ordering, and automatic external merge sort on large streams.
|
|
7
|
+
*/
|
|
8
|
+
export function sortRows(options) {
|
|
9
|
+
return externalSort(options);
|
|
10
|
+
}
|
|
11
|
+
//# sourceMappingURL=index.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/transforms/sort/index.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AAGlD,cAAc,iBAAiB,CAAC;AAChC,cAAc,oBAAoB,CAAC;AAEnC;;;GAGG;AACH,MAAM,UAAU,QAAQ,CAAC,OAAoB;IAC3C,OAAO,YAAY,CAAC,OAAO,CAAC,CAAC;AAC/B,CAAC"}
|
|
@@ -0,0 +1,6 @@
|
|
|
1
|
+
import type { TransformFunction } from "../core/types.js";
|
|
2
|
+
/**
|
|
3
|
+
* Retains and emits only the last `count` rows of a stream using a bounded O(N) Ring Buffer.
|
|
4
|
+
*/
|
|
5
|
+
export declare function tailRows(count?: number): TransformFunction;
|
|
6
|
+
//# sourceMappingURL=tail.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"tail.d.ts","sourceRoot":"","sources":["../../src/transforms/tail.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAA8B,iBAAiB,EAAE,MAAM,kBAAkB,CAAC;AAGtF;;GAEG;AACH,wBAAgB,QAAQ,CAAC,KAAK,SAAK,GAAG,iBAAiB,CAwBtD"}
|
|
@@ -0,0 +1,27 @@
|
|
|
1
|
+
import { RingBuffer } from "../utils/ring-buffer.js";
|
|
2
|
+
/**
|
|
3
|
+
* Retains and emits only the last `count` rows of a stream using a bounded O(N) Ring Buffer.
|
|
4
|
+
*/
|
|
5
|
+
export function tailRows(count = 10) {
|
|
6
|
+
const tailCount = Math.max(1, count);
|
|
7
|
+
return (stream) => {
|
|
8
|
+
return (async function* () {
|
|
9
|
+
const ringBuffer = new RingBuffer(tailCount);
|
|
10
|
+
let totalRowsSeen = 0;
|
|
11
|
+
for await (const batch of stream) {
|
|
12
|
+
for (let i = 0; i < batch.rows.length; i++) {
|
|
13
|
+
ringBuffer.push(batch.rows[i]);
|
|
14
|
+
totalRowsSeen++;
|
|
15
|
+
}
|
|
16
|
+
}
|
|
17
|
+
if (ringBuffer.size > 0) {
|
|
18
|
+
const rows = ringBuffer.toArray();
|
|
19
|
+
yield {
|
|
20
|
+
rows,
|
|
21
|
+
offset: Math.max(0, totalRowsSeen - rows.length),
|
|
22
|
+
};
|
|
23
|
+
}
|
|
24
|
+
})();
|
|
25
|
+
};
|
|
26
|
+
}
|
|
27
|
+
//# sourceMappingURL=tail.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"tail.js","sourceRoot":"","sources":["../../src/transforms/tail.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,UAAU,EAAE,MAAM,yBAAyB,CAAC;AAErD;;GAEG;AACH,MAAM,UAAU,QAAQ,CAAC,KAAK,GAAG,EAAE;IACjC,MAAM,SAAS,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,CAAC;IAErC,OAAO,CAAC,MAAkB,EAAc,EAAE;QACxC,OAAO,CAAC,KAAK,SAAS,CAAC;YACrB,MAAM,UAAU,GAAG,IAAI,UAAU,CAAM,SAAS,CAAC,CAAC;YAClD,IAAI,aAAa,GAAG,CAAC,CAAC;YAEtB,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;gBACjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;oBAC3C,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC,CAAC;oBAChC,aAAa,EAAE,CAAC;gBAClB,CAAC;YACH,CAAC;YAED,IAAI,UAAU,CAAC,IAAI,GAAG,CAAC,EAAE,CAAC;gBACxB,MAAM,IAAI,GAAG,UAAU,CAAC,OAAO,EAAE,CAAC;gBAClC,MAAM;oBACJ,IAAI;oBACJ,MAAM,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,aAAa,GAAG,IAAI,CAAC,MAAM,CAAC;iBACjD,CAAC;YACJ,CAAC;QACH,CAAC,CAAC,EAAE,CAAC;IACP,CAAC,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1,17 @@
|
|
|
1
|
+
import type { TransformFunction } from "../core/types.js";
|
|
2
|
+
import { SortKeySpec } from "./sort/comparator.js";
|
|
3
|
+
export interface TopOptions {
|
|
4
|
+
by: string | string[] | SortKeySpec[];
|
|
5
|
+
count?: number;
|
|
6
|
+
order?: "asc" | "desc";
|
|
7
|
+
smallest?: boolean;
|
|
8
|
+
nulls?: "first" | "last";
|
|
9
|
+
ignoreCase?: boolean;
|
|
10
|
+
natural?: boolean;
|
|
11
|
+
}
|
|
12
|
+
/**
|
|
13
|
+
* Retains only the Top-K (or Bottom-K) rows using a bounded O(K) Binary Heap.
|
|
14
|
+
* Time complexity: O(N * log K), Memory complexity: O(K).
|
|
15
|
+
*/
|
|
16
|
+
export declare function topRows(options: TopOptions): TransformFunction;
|
|
17
|
+
//# sourceMappingURL=top.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"top.d.ts","sourceRoot":"","sources":["../../src/transforms/top.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAA8B,iBAAiB,EAAE,MAAM,kBAAkB,CAAC;AAEtF,OAAO,EAGL,WAAW,EACZ,MAAM,sBAAsB,CAAC;AAE9B,MAAM,WAAW,UAAU;IACzB,EAAE,EAAE,MAAM,GAAG,MAAM,EAAE,GAAG,WAAW,EAAE,CAAC;IACtC,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,KAAK,CAAC,EAAE,KAAK,GAAG,MAAM,CAAC;IACvB,QAAQ,CAAC,EAAE,OAAO,CAAC;IACnB,KAAK,CAAC,EAAE,OAAO,GAAG,MAAM,CAAC;IACzB,UAAU,CAAC,EAAE,OAAO,CAAC;IACrB,OAAO,CAAC,EAAE,OAAO,CAAC;CACnB;AAED;;;GAGG;AACH,wBAAgB,OAAO,CAAC,OAAO,EAAE,UAAU,GAAG,iBAAiB,CAkD9D"}
|
|
@@ -0,0 +1,52 @@
|
|
|
1
|
+
import { Heap } from "../utils/heap.js";
|
|
2
|
+
import { createRawRowComparator, parseSortSpecs, } from "./sort/comparator.js";
|
|
3
|
+
/**
|
|
4
|
+
* Retains only the Top-K (or Bottom-K) rows using a bounded O(K) Binary Heap.
|
|
5
|
+
* Time complexity: O(N * log K), Memory complexity: O(K).
|
|
6
|
+
*/
|
|
7
|
+
export function topRows(options) {
|
|
8
|
+
const k = Math.max(1, options.count ?? 10);
|
|
9
|
+
const isSmallest = options.smallest || options.order === "asc";
|
|
10
|
+
const defaultDirection = isSmallest ? "asc" : "desc";
|
|
11
|
+
const rawSpecs = parseSortSpecs(options.by, {
|
|
12
|
+
defaultDirection,
|
|
13
|
+
nulls: options.nulls,
|
|
14
|
+
ignoreCase: options.ignoreCase,
|
|
15
|
+
natural: options.natural,
|
|
16
|
+
});
|
|
17
|
+
// Desired output comparator: desiredComparator(a, b) < 0 means 'a' comes before 'b'
|
|
18
|
+
const desiredComparator = createRawRowComparator(rawSpecs);
|
|
19
|
+
return (stream) => {
|
|
20
|
+
return (async function* () {
|
|
21
|
+
// Min-Heap where root is the LEAST preferred element among current top K
|
|
22
|
+
// If desiredComparator(a, b) > 0 ('a' is worse than 'b'), then heapComparator(a, b) < 0 (floats to root)
|
|
23
|
+
const heapComparator = (a, b) => -desiredComparator(a, b);
|
|
24
|
+
const heap = new Heap(heapComparator);
|
|
25
|
+
for await (const batch of stream) {
|
|
26
|
+
for (let i = 0; i < batch.rows.length; i++) {
|
|
27
|
+
const row = batch.rows[i];
|
|
28
|
+
if (heap.size < k) {
|
|
29
|
+
heap.push(row);
|
|
30
|
+
}
|
|
31
|
+
else {
|
|
32
|
+
const worst = heap.peek();
|
|
33
|
+
// If incoming row is more preferred than the current worst, replace it
|
|
34
|
+
if (desiredComparator(row, worst) < 0) {
|
|
35
|
+
heap.replace(row);
|
|
36
|
+
}
|
|
37
|
+
}
|
|
38
|
+
}
|
|
39
|
+
}
|
|
40
|
+
// Drain and sort results in exact desired comparator order
|
|
41
|
+
const drained = heap.drain();
|
|
42
|
+
drained.sort(desiredComparator);
|
|
43
|
+
if (drained.length > 0) {
|
|
44
|
+
yield {
|
|
45
|
+
rows: drained,
|
|
46
|
+
offset: 0,
|
|
47
|
+
};
|
|
48
|
+
}
|
|
49
|
+
})();
|
|
50
|
+
};
|
|
51
|
+
}
|
|
52
|
+
//# sourceMappingURL=top.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"top.js","sourceRoot":"","sources":["../../src/transforms/top.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,IAAI,EAAE,MAAM,kBAAkB,CAAC;AACxC,OAAO,EACL,sBAAsB,EACtB,cAAc,GAEf,MAAM,sBAAsB,CAAC;AAY9B;;;GAGG;AACH,MAAM,UAAU,OAAO,CAAC,OAAmB;IACzC,MAAM,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,OAAO,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC;IAC3C,MAAM,UAAU,GAAG,OAAO,CAAC,QAAQ,IAAI,OAAO,CAAC,KAAK,KAAK,KAAK,CAAC;IAC/D,MAAM,gBAAgB,GAAmB,UAAU,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC;IAErE,MAAM,QAAQ,GAAG,cAAc,CAAC,OAAO,CAAC,EAAE,EAAE;QAC1C,gBAAgB;QAChB,KAAK,EAAE,OAAO,CAAC,KAAK;QACpB,UAAU,EAAE,OAAO,CAAC,UAAU;QAC9B,OAAO,EAAE,OAAO,CAAC,OAAO;KACzB,CAAC,CAAC;IAEH,oFAAoF;IACpF,MAAM,iBAAiB,GAAG,sBAAsB,CAAC,QAAQ,CAAC,CAAC;IAE3D,OAAO,CAAC,MAAkB,EAAc,EAAE;QACxC,OAAO,CAAC,KAAK,SAAS,CAAC;YACrB,yEAAyE;YACzE,yGAAyG;YACzG,MAAM,cAAc,GAAG,CAAC,CAAM,EAAE,CAAM,EAAE,EAAE,CAAC,CAAC,iBAAiB,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;YACpE,MAAM,IAAI,GAAG,IAAI,IAAI,CAAM,cAAc,CAAC,CAAC;YAE3C,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;gBACjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;oBAC3C,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC;oBAE3B,IAAI,IAAI,CAAC,IAAI,GAAG,CAAC,EAAE,CAAC;wBAClB,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;oBACjB,CAAC;yBAAM,CAAC;wBACN,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,EAAG,CAAC;wBAC3B,uEAAuE;wBACvE,IAAI,iBAAiB,CAAC,GAAG,EAAE,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;4BACtC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;wBACpB,CAAC;oBACH,CAAC;gBACH,CAAC;YACH,CAAC;YAED,2DAA2D;YAC3D,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,EAAE,CAAC;YAC7B,OAAO,CAAC,IAAI,CAAC,iBAAiB,CAAC,CAAC;YAEhC,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBACvB,MAAM;oBACJ,IAAI,EAAE,OAAO;oBACb,MAAM,EAAE,CAAC;iBACV,CAAC;YACJ,CAAC;QACH,CAAC,CAAC,EAAE,CAAC;IACP,CAAC,CAAC;AACJ,CAAC"}
|
|
@@ -0,0 +1,13 @@
|
|
|
1
|
+
import type { TransformFunction } from "../core/types.js";
|
|
2
|
+
export interface UniqueOptions {
|
|
3
|
+
by?: string | string[];
|
|
4
|
+
keep?: "first" | "last";
|
|
5
|
+
memoryLimit?: number | string;
|
|
6
|
+
tempDir?: string;
|
|
7
|
+
batchSize?: number;
|
|
8
|
+
}
|
|
9
|
+
/**
|
|
10
|
+
* Deduplicates rows by key columns (or whole row) with spillable disk backup for large cardinalities.
|
|
11
|
+
*/
|
|
12
|
+
export declare function uniqueRows(options?: UniqueOptions): TransformFunction;
|
|
13
|
+
//# sourceMappingURL=unique.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"unique.d.ts","sourceRoot":"","sources":["../../src/transforms/unique.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAA8B,iBAAiB,EAAE,MAAM,kBAAkB,CAAC;AAGtF,MAAM,WAAW,aAAa;IAC5B,EAAE,CAAC,EAAE,MAAM,GAAG,MAAM,EAAE,CAAC;IACvB,IAAI,CAAC,EAAE,OAAO,GAAG,MAAM,CAAC;IACxB,WAAW,CAAC,EAAE,MAAM,GAAG,MAAM,CAAC;IAC9B,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,SAAS,CAAC,EAAE,MAAM,CAAC;CACpB;AASD;;GAEG;AACH,wBAAgB,UAAU,CAAC,OAAO,GAAE,aAAkB,GAAG,iBAAiB,CAoFzE"}
|
|
@@ -0,0 +1,91 @@
|
|
|
1
|
+
import { encodeCompositeKey } from "../diff/key.js";
|
|
2
|
+
import { SpillableKeyStore } from "../utils/keystore.js";
|
|
3
|
+
function getRowUniqueKey(row, byCols) {
|
|
4
|
+
if (byCols && byCols.length > 0) {
|
|
5
|
+
return encodeCompositeKey(row, byCols).encoded;
|
|
6
|
+
}
|
|
7
|
+
return JSON.stringify(row);
|
|
8
|
+
}
|
|
9
|
+
/**
|
|
10
|
+
* Deduplicates rows by key columns (or whole row) with spillable disk backup for large cardinalities.
|
|
11
|
+
*/
|
|
12
|
+
export function uniqueRows(options = {}) {
|
|
13
|
+
const keep = options.keep || "first";
|
|
14
|
+
const byCols = options.by
|
|
15
|
+
? (Array.isArray(options.by) ? options.by : options.by.split(",").map((s) => s.trim()).filter(Boolean))
|
|
16
|
+
: undefined;
|
|
17
|
+
const effectiveBatchSize = options.batchSize || 1000;
|
|
18
|
+
return (stream) => {
|
|
19
|
+
return (async function* () {
|
|
20
|
+
const keyStore = new SpillableKeyStore({
|
|
21
|
+
memoryLimit: options.memoryLimit,
|
|
22
|
+
tempDir: options.tempDir,
|
|
23
|
+
});
|
|
24
|
+
try {
|
|
25
|
+
if (keep === "first") {
|
|
26
|
+
// Streaming mode: Emit immediately on first encounter
|
|
27
|
+
let currentBatchRows = [];
|
|
28
|
+
let globalOffset = 0;
|
|
29
|
+
for await (const batch of stream) {
|
|
30
|
+
for (let i = 0; i < batch.rows.length; i++) {
|
|
31
|
+
const row = batch.rows[i];
|
|
32
|
+
const key = getRowUniqueKey(row, byCols);
|
|
33
|
+
const isNew = await keyStore.add(key);
|
|
34
|
+
if (isNew) {
|
|
35
|
+
currentBatchRows.push(row);
|
|
36
|
+
if (currentBatchRows.length >= effectiveBatchSize) {
|
|
37
|
+
yield {
|
|
38
|
+
rows: currentBatchRows,
|
|
39
|
+
offset: globalOffset,
|
|
40
|
+
};
|
|
41
|
+
globalOffset += currentBatchRows.length;
|
|
42
|
+
currentBatchRows = [];
|
|
43
|
+
}
|
|
44
|
+
}
|
|
45
|
+
}
|
|
46
|
+
}
|
|
47
|
+
if (currentBatchRows.length > 0) {
|
|
48
|
+
yield {
|
|
49
|
+
rows: currentBatchRows,
|
|
50
|
+
offset: globalOffset,
|
|
51
|
+
};
|
|
52
|
+
}
|
|
53
|
+
}
|
|
54
|
+
else {
|
|
55
|
+
// keep: "last" mode: Retain latest row per key
|
|
56
|
+
const latestRows = new Map();
|
|
57
|
+
let globalOffset = 0;
|
|
58
|
+
for await (const batch of stream) {
|
|
59
|
+
for (let i = 0; i < batch.rows.length; i++) {
|
|
60
|
+
const row = batch.rows[i];
|
|
61
|
+
const key = getRowUniqueKey(row, byCols);
|
|
62
|
+
latestRows.set(key, row);
|
|
63
|
+
}
|
|
64
|
+
}
|
|
65
|
+
let currentBatchRows = [];
|
|
66
|
+
for (const row of latestRows.values()) {
|
|
67
|
+
currentBatchRows.push(row);
|
|
68
|
+
if (currentBatchRows.length >= effectiveBatchSize) {
|
|
69
|
+
yield {
|
|
70
|
+
rows: currentBatchRows,
|
|
71
|
+
offset: globalOffset,
|
|
72
|
+
};
|
|
73
|
+
globalOffset += currentBatchRows.length;
|
|
74
|
+
currentBatchRows = [];
|
|
75
|
+
}
|
|
76
|
+
}
|
|
77
|
+
if (currentBatchRows.length > 0) {
|
|
78
|
+
yield {
|
|
79
|
+
rows: currentBatchRows,
|
|
80
|
+
offset: globalOffset,
|
|
81
|
+
};
|
|
82
|
+
}
|
|
83
|
+
}
|
|
84
|
+
}
|
|
85
|
+
finally {
|
|
86
|
+
await keyStore.close();
|
|
87
|
+
}
|
|
88
|
+
})();
|
|
89
|
+
};
|
|
90
|
+
}
|
|
91
|
+
//# sourceMappingURL=unique.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"unique.js","sourceRoot":"","sources":["../../src/transforms/unique.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,kBAAkB,EAAE,MAAM,gBAAgB,CAAC;AAEpD,OAAO,EAAE,iBAAiB,EAAE,MAAM,sBAAsB,CAAC;AAUzD,SAAS,eAAe,CAAC,GAAQ,EAAE,MAAiB;IAClD,IAAI,MAAM,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAChC,OAAO,kBAAkB,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC,OAAO,CAAC;IACjD,CAAC;IACD,OAAO,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC;AAC7B,CAAC;AAED;;GAEG;AACH,MAAM,UAAU,UAAU,CAAC,UAAyB,EAAE;IACpD,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,IAAI,OAAO,CAAC;IACrC,MAAM,MAAM,GAAG,OAAO,CAAC,EAAE;QACvB,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QACvG,CAAC,CAAC,SAAS,CAAC;IACd,MAAM,kBAAkB,GAAG,OAAO,CAAC,SAAS,IAAI,IAAI,CAAC;IAErD,OAAO,CAAC,MAAkB,EAAc,EAAE;QACxC,OAAO,CAAC,KAAK,SAAS,CAAC;YACrB,MAAM,QAAQ,GAAG,IAAI,iBAAiB,CAAC;gBACrC,WAAW,EAAE,OAAO,CAAC,WAAW;gBAChC,OAAO,EAAE,OAAO,CAAC,OAAO;aACzB,CAAC,CAAC;YAEH,IAAI,CAAC;gBACH,IAAI,IAAI,KAAK,OAAO,EAAE,CAAC;oBACrB,sDAAsD;oBACtD,IAAI,gBAAgB,GAAU,EAAE,CAAC;oBACjC,IAAI,YAAY,GAAG,CAAC,CAAC;oBAErB,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;wBACjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;4BAC3C,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC;4BAC3B,MAAM,GAAG,GAAG,eAAe,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;4BAEzC,MAAM,KAAK,GAAG,MAAM,QAAQ,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;4BACtC,IAAI,KAAK,EAAE,CAAC;gCACV,gBAAgB,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;gCAC3B,IAAI,gBAAgB,CAAC,MAAM,IAAI,kBAAkB,EAAE,CAAC;oCAClD,MAAM;wCACJ,IAAI,EAAE,gBAAgB;wCACtB,MAAM,EAAE,YAAY;qCACrB,CAAC;oCACF,YAAY,IAAI,gBAAgB,CAAC,MAAM,CAAC;oCACxC,gBAAgB,GAAG,EAAE,CAAC;gCACxB,CAAC;4BACH,CAAC;wBACH,CAAC;oBACH,CAAC;oBAED,IAAI,gBAAgB,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;wBAChC,MAAM;4BACJ,IAAI,EAAE,gBAAgB;4BACtB,MAAM,EAAE,YAAY;yBACrB,CAAC;oBACJ,CAAC;gBACH,CAAC;qBAAM,CAAC;oBACN,+CAA+C;oBAC/C,MAAM,UAAU,GAAG,IAAI,GAAG,EAAe,CAAC;oBAC1C,IAAI,YAAY,GAAG,CAAC,CAAC;oBAErB,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,EAAE,CAAC;wBACjC,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;4BAC3C,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,CAAE,CAAC;4BAC3B,MAAM,GAAG,GAAG,eAAe,CAAC,GAAG,EAAE,MAAM,CAAC,CAAC;4BACzC,UAAU,CAAC,GAAG,CAAC,GAAG,EAAE,GAAG,CAAC,CAAC;wBAC3B,CAAC;oBACH,CAAC;oBAED,IAAI,gBAAgB,GAAU,EAAE,CAAC;oBACjC,KAAK,MAAM,GAAG,IAAI,UAAU,CAAC,MAAM,EAAE,EAAE,CAAC;wBACtC,gBAAgB,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;wBAC3B,IAAI,gBAAgB,CAAC,MAAM,IAAI,kBAAkB,EAAE,CAAC;4BAClD,MAAM;gCACJ,IAAI,EAAE,gBAAgB;gCACtB,MAAM,EAAE,YAAY;6BACrB,CAAC;4BACF,YAAY,IAAI,gBAAgB,CAAC,MAAM,CAAC;4BACxC,gBAAgB,GAAG,EAAE,CAAC;wBACxB,CAAC;oBACH,CAAC;oBAED,IAAI,gBAAgB,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;wBAChC,MAAM;4BACJ,IAAI,EAAE,gBAAgB;4BACtB,MAAM,EAAE,YAAY;yBACrB,CAAC;oBACJ,CAAC;gBACH,CAAC;YACH,CAAC;oBAAS,CAAC;gBACT,MAAM,QAAQ,CAAC,KAAK,EAAE,CAAC;YACzB,CAAC;QACH,CAAC,CAAC,EAAE,CAAC;IACP,CAAC,CAAC;AACJ,CAAC"}
|
|
@@ -1,6 +1,7 @@
|
|
|
1
1
|
export declare function formatNumber(num: number | null | undefined): string;
|
|
2
2
|
export declare function formatDecimal(num: number | null | undefined, digits?: number): string;
|
|
3
3
|
export declare function formatBytes(bytes: number | null | undefined): string;
|
|
4
|
+
export declare function parseBytes(input: string | number | null | undefined): number;
|
|
4
5
|
export declare function formatTable(headers: string[], rows: string[][], alignments?: Array<"left" | "right">): string;
|
|
5
6
|
export declare function logMemoryDebug(): void;
|
|
6
7
|
//# sourceMappingURL=formatting.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"formatting.d.ts","sourceRoot":"","sources":["../../src/utils/formatting.ts"],"names":[],"mappings":"AAAA,wBAAgB,YAAY,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI,GAAG,SAAS,GAAG,MAAM,CAKnE;AAED,wBAAgB,aAAa,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI,GAAG,SAAS,EAAE,MAAM,SAAI,GAAG,MAAM,CAQhF;AAED,wBAAgB,WAAW,CAAC,KAAK,EAAE,MAAM,GAAG,IAAI,GAAG,SAAS,GAAG,MAAM,
|
|
1
|
+
{"version":3,"file":"formatting.d.ts","sourceRoot":"","sources":["../../src/utils/formatting.ts"],"names":[],"mappings":"AAAA,wBAAgB,YAAY,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI,GAAG,SAAS,GAAG,MAAM,CAKnE;AAED,wBAAgB,aAAa,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI,GAAG,SAAS,EAAE,MAAM,SAAI,GAAG,MAAM,CAQhF;AAED,wBAAgB,WAAW,CAAC,KAAK,EAAE,MAAM,GAAG,IAAI,GAAG,SAAS,GAAG,MAAM,CAQpE;AAED,wBAAgB,UAAU,CAAC,KAAK,EAAE,MAAM,GAAG,MAAM,GAAG,IAAI,GAAG,SAAS,GAAG,MAAM,CAyB5E;AAED,wBAAgB,WAAW,CACzB,OAAO,EAAE,MAAM,EAAE,EACjB,IAAI,EAAE,MAAM,EAAE,EAAE,EAChB,UAAU,CAAC,EAAE,KAAK,CAAC,MAAM,GAAG,OAAO,CAAC,GACnC,MAAM,CAoCR;AAED,wBAAgB,cAAc,IAAI,IAAI,CAOrC"}
|