@loaders.gl/parquet 4.0.0-alpha.22 → 4.0.0-alpha.24

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.
Files changed (104) hide show
  1. package/dist/dist.min.js +17 -17
  2. package/dist/dist.min.js.map +3 -3
  3. package/dist/es5/index.js +0 -3
  4. package/dist/es5/index.js.map +1 -1
  5. package/dist/es5/parquet-loader.js +1 -1
  6. package/dist/es5/parquet-wasm-loader.js +1 -1
  7. package/dist/es5/parquet-wasm-writer.js +1 -1
  8. package/dist/es5/parquet-writer.js +2 -3
  9. package/dist/es5/parquet-writer.js.map +1 -1
  10. package/dist/esm/index.js +0 -1
  11. package/dist/esm/index.js.map +1 -1
  12. package/dist/esm/parquet-loader.js +1 -1
  13. package/dist/esm/parquet-wasm-loader.js +1 -1
  14. package/dist/esm/parquet-wasm-writer.js +1 -1
  15. package/dist/esm/parquet-writer.js +2 -3
  16. package/dist/esm/parquet-writer.js.map +1 -1
  17. package/dist/index.d.ts +0 -1
  18. package/dist/index.d.ts.map +1 -1
  19. package/dist/parquet-worker.js +17 -17
  20. package/dist/parquet-worker.js.map +3 -3
  21. package/dist/parquet-writer.d.ts.map +1 -1
  22. package/package.json +11 -9
  23. package/src/index.ts +0 -3
  24. package/src/parquet-writer.ts +1 -3
  25. package/dist/bundle.js +0 -5
  26. package/dist/constants.js +0 -18
  27. package/dist/index.js +0 -58
  28. package/dist/lib/arrow/convert-columns-to-row-group.js +0 -1
  29. package/dist/lib/arrow/convert-row-group-to-columns.js +0 -12
  30. package/dist/lib/arrow/convert-schema-from-parquet.js +0 -86
  31. package/dist/lib/arrow/convert-schema-to-parquet.js +0 -71
  32. package/dist/lib/geo/decode-geo-metadata.js +0 -77
  33. package/dist/lib/geo/geoparquet-schema.js +0 -69
  34. package/dist/lib/parsers/parse-parquet-to-columns.js +0 -46
  35. package/dist/lib/parsers/parse-parquet-to-rows.js +0 -37
  36. package/dist/lib/wasm/encode-parquet-wasm.js +0 -30
  37. package/dist/lib/wasm/load-wasm/index.js +0 -5
  38. package/dist/lib/wasm/load-wasm/load-wasm-browser.js +0 -38
  39. package/dist/lib/wasm/load-wasm/load-wasm-node.js +0 -31
  40. package/dist/lib/wasm/parse-parquet-wasm.js +0 -27
  41. package/dist/parquet-loader.js +0 -41
  42. package/dist/parquet-wasm-loader.js +0 -26
  43. package/dist/parquet-wasm-writer.js +0 -24
  44. package/dist/parquet-writer.js +0 -22
  45. package/dist/parquetjs/codecs/declare.js +0 -2
  46. package/dist/parquetjs/codecs/dictionary.js +0 -14
  47. package/dist/parquetjs/codecs/index.js +0 -55
  48. package/dist/parquetjs/codecs/plain.js +0 -211
  49. package/dist/parquetjs/codecs/rle.js +0 -145
  50. package/dist/parquetjs/compression.js +0 -183
  51. package/dist/parquetjs/encoder/parquet-encoder.js +0 -484
  52. package/dist/parquetjs/parquet-thrift/BoundaryOrder.js +0 -15
  53. package/dist/parquetjs/parquet-thrift/BsonType.js +0 -62
  54. package/dist/parquetjs/parquet-thrift/ColumnChunk.js +0 -211
  55. package/dist/parquetjs/parquet-thrift/ColumnIndex.js +0 -217
  56. package/dist/parquetjs/parquet-thrift/ColumnMetaData.js +0 -402
  57. package/dist/parquetjs/parquet-thrift/ColumnOrder.js +0 -108
  58. package/dist/parquetjs/parquet-thrift/CompressionCodec.js +0 -20
  59. package/dist/parquetjs/parquet-thrift/ConvertedType.js +0 -34
  60. package/dist/parquetjs/parquet-thrift/DataPageHeader.js +0 -170
  61. package/dist/parquetjs/parquet-thrift/DataPageHeaderV2.js +0 -230
  62. package/dist/parquetjs/parquet-thrift/DateType.js +0 -62
  63. package/dist/parquetjs/parquet-thrift/DecimalType.js +0 -109
  64. package/dist/parquetjs/parquet-thrift/DictionaryPageHeader.js +0 -126
  65. package/dist/parquetjs/parquet-thrift/Encoding.js +0 -20
  66. package/dist/parquetjs/parquet-thrift/EnumType.js +0 -62
  67. package/dist/parquetjs/parquet-thrift/FieldRepetitionType.js +0 -15
  68. package/dist/parquetjs/parquet-thrift/FileMetaData.js +0 -260
  69. package/dist/parquetjs/parquet-thrift/IndexPageHeader.js +0 -62
  70. package/dist/parquetjs/parquet-thrift/IntType.js +0 -109
  71. package/dist/parquetjs/parquet-thrift/JsonType.js +0 -62
  72. package/dist/parquetjs/parquet-thrift/KeyValue.js +0 -106
  73. package/dist/parquetjs/parquet-thrift/ListType.js +0 -62
  74. package/dist/parquetjs/parquet-thrift/LogicalType.js +0 -384
  75. package/dist/parquetjs/parquet-thrift/MapType.js +0 -62
  76. package/dist/parquetjs/parquet-thrift/MicroSeconds.js +0 -62
  77. package/dist/parquetjs/parquet-thrift/MilliSeconds.js +0 -62
  78. package/dist/parquetjs/parquet-thrift/NullType.js +0 -62
  79. package/dist/parquetjs/parquet-thrift/OffsetIndex.js +0 -101
  80. package/dist/parquetjs/parquet-thrift/PageEncodingStats.js +0 -131
  81. package/dist/parquetjs/parquet-thrift/PageHeader.js +0 -220
  82. package/dist/parquetjs/parquet-thrift/PageLocation.js +0 -145
  83. package/dist/parquetjs/parquet-thrift/PageType.js +0 -16
  84. package/dist/parquetjs/parquet-thrift/RowGroup.js +0 -186
  85. package/dist/parquetjs/parquet-thrift/SchemaElement.js +0 -243
  86. package/dist/parquetjs/parquet-thrift/SortingColumn.js +0 -131
  87. package/dist/parquetjs/parquet-thrift/Statistics.js +0 -180
  88. package/dist/parquetjs/parquet-thrift/StringType.js +0 -62
  89. package/dist/parquetjs/parquet-thrift/TimeType.js +0 -110
  90. package/dist/parquetjs/parquet-thrift/TimeUnit.js +0 -131
  91. package/dist/parquetjs/parquet-thrift/TimestampType.js +0 -110
  92. package/dist/parquetjs/parquet-thrift/Type.js +0 -20
  93. package/dist/parquetjs/parquet-thrift/TypeDefinedOrder.js +0 -62
  94. package/dist/parquetjs/parquet-thrift/UUIDType.js +0 -62
  95. package/dist/parquetjs/parquet-thrift/index.js +0 -65
  96. package/dist/parquetjs/parser/decoders.js +0 -318
  97. package/dist/parquetjs/parser/parquet-reader.js +0 -200
  98. package/dist/parquetjs/schema/declare.js +0 -12
  99. package/dist/parquetjs/schema/schema.js +0 -162
  100. package/dist/parquetjs/schema/shred.js +0 -355
  101. package/dist/parquetjs/schema/types.js +0 -416
  102. package/dist/parquetjs/utils/file-utils.js +0 -43
  103. package/dist/parquetjs/utils/read-utils.js +0 -109
  104. package/dist/workers/parquet-worker.js +0 -5
@@ -1,110 +0,0 @@
1
- "use strict";
2
- var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
3
- if (k2 === undefined) k2 = k;
4
- var desc = Object.getOwnPropertyDescriptor(m, k);
5
- if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
6
- desc = { enumerable: true, get: function() { return m[k]; } };
7
- }
8
- Object.defineProperty(o, k2, desc);
9
- }) : (function(o, m, k, k2) {
10
- if (k2 === undefined) k2 = k;
11
- o[k2] = m[k];
12
- }));
13
- var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
14
- Object.defineProperty(o, "default", { enumerable: true, value: v });
15
- }) : function(o, v) {
16
- o["default"] = v;
17
- });
18
- var __importStar = (this && this.__importStar) || function (mod) {
19
- if (mod && mod.__esModule) return mod;
20
- var result = {};
21
- if (mod != null) for (var k in mod) if (k !== "default" && Object.prototype.hasOwnProperty.call(mod, k)) __createBinding(result, mod, k);
22
- __setModuleDefault(result, mod);
23
- return result;
24
- };
25
- Object.defineProperty(exports, "__esModule", { value: true });
26
- exports.TimestampType = void 0;
27
- /* tslint:disable */
28
- /* eslint-disable */
29
- /*
30
- * Autogenerated by @creditkarma/thrift-typescript v3.7.2
31
- * DO NOT EDIT UNLESS YOU ARE SURE THAT YOU KNOW WHAT YOU ARE DOING
32
- */
33
- const thrift = __importStar(require("thrift"));
34
- const TimeUnit = __importStar(require("./TimeUnit"));
35
- class TimestampType {
36
- constructor(args) {
37
- if (args != null && args.isAdjustedToUTC != null) {
38
- this.isAdjustedToUTC = args.isAdjustedToUTC;
39
- }
40
- else {
41
- throw new thrift.Thrift.TProtocolException(thrift.Thrift.TProtocolExceptionType.UNKNOWN, 'Required field[isAdjustedToUTC] is unset!');
42
- }
43
- if (args != null && args.unit != null) {
44
- this.unit = args.unit;
45
- }
46
- else {
47
- throw new thrift.Thrift.TProtocolException(thrift.Thrift.TProtocolExceptionType.UNKNOWN, 'Required field[unit] is unset!');
48
- }
49
- }
50
- write(output) {
51
- output.writeStructBegin('TimestampType');
52
- if (this.isAdjustedToUTC != null) {
53
- output.writeFieldBegin('isAdjustedToUTC', thrift.Thrift.Type.BOOL, 1);
54
- output.writeBool(this.isAdjustedToUTC);
55
- output.writeFieldEnd();
56
- }
57
- if (this.unit != null) {
58
- output.writeFieldBegin('unit', thrift.Thrift.Type.STRUCT, 2);
59
- this.unit.write(output);
60
- output.writeFieldEnd();
61
- }
62
- output.writeFieldStop();
63
- output.writeStructEnd();
64
- return;
65
- }
66
- static read(input) {
67
- input.readStructBegin();
68
- let _args = {};
69
- while (true) {
70
- const ret = input.readFieldBegin();
71
- const fieldType = ret.ftype;
72
- const fieldId = ret.fid;
73
- if (fieldType === thrift.Thrift.Type.STOP) {
74
- break;
75
- }
76
- switch (fieldId) {
77
- case 1:
78
- if (fieldType === thrift.Thrift.Type.BOOL) {
79
- const value_1 = input.readBool();
80
- _args.isAdjustedToUTC = value_1;
81
- }
82
- else {
83
- input.skip(fieldType);
84
- }
85
- break;
86
- case 2:
87
- if (fieldType === thrift.Thrift.Type.STRUCT) {
88
- const value_2 = TimeUnit.TimeUnit.read(input);
89
- _args.unit = value_2;
90
- }
91
- else {
92
- input.skip(fieldType);
93
- }
94
- break;
95
- default: {
96
- input.skip(fieldType);
97
- }
98
- }
99
- input.readFieldEnd();
100
- }
101
- input.readStructEnd();
102
- if (_args.isAdjustedToUTC !== undefined && _args.unit !== undefined) {
103
- return new TimestampType(_args);
104
- }
105
- else {
106
- throw new thrift.Thrift.TProtocolException(thrift.Thrift.TProtocolExceptionType.UNKNOWN, 'Unable to read TimestampType from input');
107
- }
108
- }
109
- }
110
- exports.TimestampType = TimestampType;
@@ -1,20 +0,0 @@
1
- "use strict";
2
- Object.defineProperty(exports, "__esModule", { value: true });
3
- exports.Type = void 0;
4
- /* tslint:disable */
5
- /* eslint-disable */
6
- /*
7
- * Autogenerated by @creditkarma/thrift-typescript v3.7.2
8
- * DO NOT EDIT UNLESS YOU ARE SURE THAT YOU KNOW WHAT YOU ARE DOING
9
- */
10
- var Type;
11
- (function (Type) {
12
- Type[Type["BOOLEAN"] = 0] = "BOOLEAN";
13
- Type[Type["INT32"] = 1] = "INT32";
14
- Type[Type["INT64"] = 2] = "INT64";
15
- Type[Type["INT96"] = 3] = "INT96";
16
- Type[Type["FLOAT"] = 4] = "FLOAT";
17
- Type[Type["DOUBLE"] = 5] = "DOUBLE";
18
- Type[Type["BYTE_ARRAY"] = 6] = "BYTE_ARRAY";
19
- Type[Type["FIXED_LEN_BYTE_ARRAY"] = 7] = "FIXED_LEN_BYTE_ARRAY";
20
- })(Type || (exports.Type = Type = {}));
@@ -1,62 +0,0 @@
1
- "use strict";
2
- var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
3
- if (k2 === undefined) k2 = k;
4
- var desc = Object.getOwnPropertyDescriptor(m, k);
5
- if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
6
- desc = { enumerable: true, get: function() { return m[k]; } };
7
- }
8
- Object.defineProperty(o, k2, desc);
9
- }) : (function(o, m, k, k2) {
10
- if (k2 === undefined) k2 = k;
11
- o[k2] = m[k];
12
- }));
13
- var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
14
- Object.defineProperty(o, "default", { enumerable: true, value: v });
15
- }) : function(o, v) {
16
- o["default"] = v;
17
- });
18
- var __importStar = (this && this.__importStar) || function (mod) {
19
- if (mod && mod.__esModule) return mod;
20
- var result = {};
21
- if (mod != null) for (var k in mod) if (k !== "default" && Object.prototype.hasOwnProperty.call(mod, k)) __createBinding(result, mod, k);
22
- __setModuleDefault(result, mod);
23
- return result;
24
- };
25
- Object.defineProperty(exports, "__esModule", { value: true });
26
- exports.TypeDefinedOrder = void 0;
27
- /* tslint:disable */
28
- /* eslint-disable */
29
- /*
30
- * Autogenerated by @creditkarma/thrift-typescript v3.7.2
31
- * DO NOT EDIT UNLESS YOU ARE SURE THAT YOU KNOW WHAT YOU ARE DOING
32
- */
33
- const thrift = __importStar(require("thrift"));
34
- class TypeDefinedOrder {
35
- constructor() { }
36
- write(output) {
37
- output.writeStructBegin('TypeDefinedOrder');
38
- output.writeFieldStop();
39
- output.writeStructEnd();
40
- return;
41
- }
42
- static read(input) {
43
- input.readStructBegin();
44
- while (true) {
45
- const ret = input.readFieldBegin();
46
- const fieldType = ret.ftype;
47
- const fieldId = ret.fid;
48
- if (fieldType === thrift.Thrift.Type.STOP) {
49
- break;
50
- }
51
- switch (fieldId) {
52
- default: {
53
- input.skip(fieldType);
54
- }
55
- }
56
- input.readFieldEnd();
57
- }
58
- input.readStructEnd();
59
- return new TypeDefinedOrder();
60
- }
61
- }
62
- exports.TypeDefinedOrder = TypeDefinedOrder;
@@ -1,62 +0,0 @@
1
- "use strict";
2
- var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
3
- if (k2 === undefined) k2 = k;
4
- var desc = Object.getOwnPropertyDescriptor(m, k);
5
- if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
6
- desc = { enumerable: true, get: function() { return m[k]; } };
7
- }
8
- Object.defineProperty(o, k2, desc);
9
- }) : (function(o, m, k, k2) {
10
- if (k2 === undefined) k2 = k;
11
- o[k2] = m[k];
12
- }));
13
- var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
14
- Object.defineProperty(o, "default", { enumerable: true, value: v });
15
- }) : function(o, v) {
16
- o["default"] = v;
17
- });
18
- var __importStar = (this && this.__importStar) || function (mod) {
19
- if (mod && mod.__esModule) return mod;
20
- var result = {};
21
- if (mod != null) for (var k in mod) if (k !== "default" && Object.prototype.hasOwnProperty.call(mod, k)) __createBinding(result, mod, k);
22
- __setModuleDefault(result, mod);
23
- return result;
24
- };
25
- Object.defineProperty(exports, "__esModule", { value: true });
26
- exports.UUIDType = void 0;
27
- /* tslint:disable */
28
- /* eslint-disable */
29
- /*
30
- * Autogenerated by @creditkarma/thrift-typescript v3.7.2
31
- * DO NOT EDIT UNLESS YOU ARE SURE THAT YOU KNOW WHAT YOU ARE DOING
32
- */
33
- const thrift = __importStar(require("thrift"));
34
- class UUIDType {
35
- constructor() { }
36
- write(output) {
37
- output.writeStructBegin('UUIDType');
38
- output.writeFieldStop();
39
- output.writeStructEnd();
40
- return;
41
- }
42
- static read(input) {
43
- input.readStructBegin();
44
- while (true) {
45
- const ret = input.readFieldBegin();
46
- const fieldType = ret.ftype;
47
- const fieldId = ret.fid;
48
- if (fieldType === thrift.Thrift.Type.STOP) {
49
- break;
50
- }
51
- switch (fieldId) {
52
- default: {
53
- input.skip(fieldType);
54
- }
55
- }
56
- input.readFieldEnd();
57
- }
58
- input.readStructEnd();
59
- return new UUIDType();
60
- }
61
- }
62
- exports.UUIDType = UUIDType;
@@ -1,65 +0,0 @@
1
- "use strict";
2
- var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
3
- if (k2 === undefined) k2 = k;
4
- var desc = Object.getOwnPropertyDescriptor(m, k);
5
- if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
6
- desc = { enumerable: true, get: function() { return m[k]; } };
7
- }
8
- Object.defineProperty(o, k2, desc);
9
- }) : (function(o, m, k, k2) {
10
- if (k2 === undefined) k2 = k;
11
- o[k2] = m[k];
12
- }));
13
- var __exportStar = (this && this.__exportStar) || function(m, exports) {
14
- for (var p in m) if (p !== "default" && !Object.prototype.hasOwnProperty.call(exports, p)) __createBinding(exports, m, p);
15
- };
16
- Object.defineProperty(exports, "__esModule", { value: true });
17
- /* tslint:disable */
18
- /* eslint-disable */
19
- /*
20
- * Autogenerated by @creditkarma/thrift-typescript v3.7.2
21
- * DO NOT EDIT UNLESS YOU ARE SURE THAT YOU KNOW WHAT YOU ARE DOING
22
- */
23
- __exportStar(require("./Type"), exports);
24
- __exportStar(require("./ConvertedType"), exports);
25
- __exportStar(require("./FieldRepetitionType"), exports);
26
- __exportStar(require("./Encoding"), exports);
27
- __exportStar(require("./CompressionCodec"), exports);
28
- __exportStar(require("./PageType"), exports);
29
- __exportStar(require("./BoundaryOrder"), exports);
30
- __exportStar(require("./Statistics"), exports);
31
- __exportStar(require("./StringType"), exports);
32
- __exportStar(require("./UUIDType"), exports);
33
- __exportStar(require("./MapType"), exports);
34
- __exportStar(require("./ListType"), exports);
35
- __exportStar(require("./EnumType"), exports);
36
- __exportStar(require("./DateType"), exports);
37
- __exportStar(require("./NullType"), exports);
38
- __exportStar(require("./DecimalType"), exports);
39
- __exportStar(require("./MilliSeconds"), exports);
40
- __exportStar(require("./MicroSeconds"), exports);
41
- __exportStar(require("./TimestampType"), exports);
42
- __exportStar(require("./TimeType"), exports);
43
- __exportStar(require("./IntType"), exports);
44
- __exportStar(require("./JsonType"), exports);
45
- __exportStar(require("./BsonType"), exports);
46
- __exportStar(require("./SchemaElement"), exports);
47
- __exportStar(require("./DataPageHeader"), exports);
48
- __exportStar(require("./IndexPageHeader"), exports);
49
- __exportStar(require("./DictionaryPageHeader"), exports);
50
- __exportStar(require("./DataPageHeaderV2"), exports);
51
- __exportStar(require("./PageHeader"), exports);
52
- __exportStar(require("./KeyValue"), exports);
53
- __exportStar(require("./SortingColumn"), exports);
54
- __exportStar(require("./PageEncodingStats"), exports);
55
- __exportStar(require("./ColumnMetaData"), exports);
56
- __exportStar(require("./ColumnChunk"), exports);
57
- __exportStar(require("./RowGroup"), exports);
58
- __exportStar(require("./TypeDefinedOrder"), exports);
59
- __exportStar(require("./PageLocation"), exports);
60
- __exportStar(require("./OffsetIndex"), exports);
61
- __exportStar(require("./ColumnIndex"), exports);
62
- __exportStar(require("./FileMetaData"), exports);
63
- __exportStar(require("./TimeUnit"), exports);
64
- __exportStar(require("./LogicalType"), exports);
65
- __exportStar(require("./ColumnOrder"), exports);
@@ -1,318 +0,0 @@
1
- "use strict";
2
- Object.defineProperty(exports, "__esModule", { value: true });
3
- exports.decodeSchema = exports.decodePage = exports.decodeDataPages = void 0;
4
- const codecs_1 = require("../codecs");
5
- const parquet_thrift_1 = require("../parquet-thrift");
6
- const compression_1 = require("../compression");
7
- const constants_1 = require("../../constants");
8
- const read_utils_1 = require("../utils/read-utils");
9
- /**
10
- * Decode data pages
11
- * @param buffer - input data
12
- * @param column - parquet column
13
- * @param compression - compression type
14
- * @returns parquet data page data
15
- */
16
- async function decodeDataPages(buffer, options) {
17
- const cursor = {
18
- buffer,
19
- offset: 0,
20
- size: buffer.length
21
- };
22
- const data = {
23
- rlevels: [],
24
- dlevels: [],
25
- values: [],
26
- pageHeaders: [],
27
- count: 0
28
- };
29
- let dictionary = options.dictionary || [];
30
- while (
31
- // @ts-ignore size can be undefined
32
- cursor.offset < cursor.size &&
33
- (!options.numValues || data.dlevels.length < Number(options.numValues))) {
34
- // Looks like we have to decode these in sequence due to cursor updates?
35
- const page = await decodePage(cursor, options);
36
- if (page.dictionary) {
37
- dictionary = page.dictionary;
38
- // eslint-disable-next-line no-continue
39
- continue;
40
- }
41
- if (dictionary.length) {
42
- // eslint-disable-next-line no-loop-func
43
- page.values = page.values.map((value) => dictionary[value]);
44
- }
45
- for (let index = 0; index < page.rlevels.length; index++) {
46
- data.rlevels.push(page.rlevels[index]);
47
- data.dlevels.push(page.dlevels[index]);
48
- const value = page.values[index];
49
- if (value !== undefined) {
50
- data.values.push(value);
51
- }
52
- }
53
- data.count += page.count;
54
- data.pageHeaders.push(page.pageHeader);
55
- }
56
- return data;
57
- }
58
- exports.decodeDataPages = decodeDataPages;
59
- /**
60
- * Decode parquet page based on page type
61
- * @param cursor
62
- * @param options
63
- */
64
- async function decodePage(cursor, options) {
65
- let page;
66
- const { pageHeader, length } = (0, read_utils_1.decodePageHeader)(cursor.buffer, cursor.offset);
67
- cursor.offset += length;
68
- const pageType = (0, read_utils_1.getThriftEnum)(parquet_thrift_1.PageType, pageHeader.type);
69
- switch (pageType) {
70
- case 'DATA_PAGE':
71
- page = await decodeDataPage(cursor, pageHeader, options);
72
- break;
73
- case 'DATA_PAGE_V2':
74
- page = await decodeDataPageV2(cursor, pageHeader, options);
75
- break;
76
- case 'DICTIONARY_PAGE':
77
- page = {
78
- dictionary: await decodeDictionaryPage(cursor, pageHeader, options),
79
- pageHeader
80
- };
81
- break;
82
- default:
83
- throw new Error(`invalid page type: ${pageType}`);
84
- }
85
- return page;
86
- }
87
- exports.decodePage = decodePage;
88
- /**
89
- * Decode parquet schema
90
- * @param schemaElements input schema elements data
91
- * @param offset offset to read from
92
- * @param len length of data
93
- * @returns result.offset
94
- * result.next - offset at the end of function
95
- * result.schema - schema read from the input data
96
- * @todo output offset is the same as input - possibly excess output field
97
- */
98
- function decodeSchema(schemaElements, offset, len) {
99
- const schema = {};
100
- let next = offset;
101
- for (let i = 0; i < len; i++) {
102
- const schemaElement = schemaElements[next];
103
- const repetitionType = next > 0 ? (0, read_utils_1.getThriftEnum)(parquet_thrift_1.FieldRepetitionType, schemaElement.repetition_type) : 'ROOT';
104
- let optional = false;
105
- let repeated = false;
106
- switch (repetitionType) {
107
- case 'REQUIRED':
108
- break;
109
- case 'OPTIONAL':
110
- optional = true;
111
- break;
112
- case 'REPEATED':
113
- repeated = true;
114
- break;
115
- default:
116
- throw new Error('parquet: unknown repetition type');
117
- }
118
- if (schemaElement.num_children > 0) {
119
- const res = decodeSchema(schemaElements, next + 1, schemaElement.num_children);
120
- next = res.next;
121
- schema[schemaElement.name] = {
122
- // type: undefined,
123
- optional,
124
- repeated,
125
- fields: res.schema
126
- };
127
- }
128
- else {
129
- const type = (0, read_utils_1.getThriftEnum)(parquet_thrift_1.Type, schemaElement.type);
130
- let logicalType = type;
131
- if (schemaElement.converted_type) {
132
- logicalType = (0, read_utils_1.getThriftEnum)(parquet_thrift_1.ConvertedType, schemaElement.converted_type);
133
- }
134
- switch (logicalType) {
135
- case 'DECIMAL':
136
- logicalType = `${logicalType}_${type}`;
137
- break;
138
- default:
139
- }
140
- schema[schemaElement.name] = {
141
- type: logicalType,
142
- typeLength: schemaElement.type_length,
143
- presision: schemaElement.precision,
144
- scale: schemaElement.scale,
145
- optional,
146
- repeated
147
- };
148
- next++;
149
- }
150
- }
151
- return { schema, offset, next };
152
- }
153
- exports.decodeSchema = decodeSchema;
154
- /**
155
- * Decode a consecutive array of data using one of the parquet encodings
156
- */
157
- function decodeValues(type, encoding, cursor, count, opts) {
158
- if (!(encoding in codecs_1.PARQUET_CODECS)) {
159
- throw new Error(`invalid encoding: ${encoding}`);
160
- }
161
- return codecs_1.PARQUET_CODECS[encoding].decodeValues(type, cursor, count, opts);
162
- }
163
- /**
164
- * Do decoding of parquet dataPage from column chunk
165
- * @param cursor
166
- * @param header
167
- * @param options
168
- */
169
- async function decodeDataPage(cursor, header, options) {
170
- const cursorEnd = cursor.offset + header.compressed_page_size;
171
- const valueCount = header.data_page_header?.num_values;
172
- /* uncompress page */
173
- let dataCursor = cursor;
174
- if (options.compression !== 'UNCOMPRESSED') {
175
- const valuesBuf = await (0, compression_1.decompress)(options.compression, cursor.buffer.slice(cursor.offset, cursorEnd), header.uncompressed_page_size);
176
- dataCursor = {
177
- buffer: valuesBuf,
178
- offset: 0,
179
- size: valuesBuf.length
180
- };
181
- cursor.offset = cursorEnd;
182
- }
183
- /* read repetition levels */
184
- const rLevelEncoding = (0, read_utils_1.getThriftEnum)(parquet_thrift_1.Encoding, header.data_page_header?.repetition_level_encoding);
185
- // tslint:disable-next-line:prefer-array-literal
186
- let rLevels = new Array(valueCount);
187
- if (options.column.rLevelMax > 0) {
188
- rLevels = decodeValues(constants_1.PARQUET_RDLVL_TYPE, rLevelEncoding, dataCursor, valueCount, {
189
- bitWidth: (0, read_utils_1.getBitWidth)(options.column.rLevelMax),
190
- disableEnvelope: false
191
- // column: opts.column
192
- });
193
- }
194
- else {
195
- rLevels.fill(0);
196
- }
197
- /* read definition levels */
198
- const dLevelEncoding = (0, read_utils_1.getThriftEnum)(parquet_thrift_1.Encoding, header.data_page_header?.definition_level_encoding);
199
- // tslint:disable-next-line:prefer-array-literal
200
- let dLevels = new Array(valueCount);
201
- if (options.column.dLevelMax > 0) {
202
- dLevels = decodeValues(constants_1.PARQUET_RDLVL_TYPE, dLevelEncoding, dataCursor, valueCount, {
203
- bitWidth: (0, read_utils_1.getBitWidth)(options.column.dLevelMax),
204
- disableEnvelope: false
205
- // column: opts.column
206
- });
207
- }
208
- else {
209
- dLevels.fill(0);
210
- }
211
- let valueCountNonNull = 0;
212
- for (const dlvl of dLevels) {
213
- if (dlvl === options.column.dLevelMax) {
214
- valueCountNonNull++;
215
- }
216
- }
217
- /* read values */
218
- const valueEncoding = (0, read_utils_1.getThriftEnum)(parquet_thrift_1.Encoding, header.data_page_header?.encoding);
219
- const decodeOptions = {
220
- typeLength: options.column.typeLength,
221
- bitWidth: options.column.typeLength
222
- };
223
- const values = decodeValues(options.column.primitiveType, valueEncoding, dataCursor, valueCountNonNull, decodeOptions);
224
- return {
225
- dlevels: dLevels,
226
- rlevels: rLevels,
227
- values,
228
- count: valueCount,
229
- pageHeader: header
230
- };
231
- }
232
- /**
233
- * Do decoding of parquet dataPage in version 2 from column chunk
234
- * @param cursor
235
- * @param header
236
- * @param opts
237
- * @returns
238
- */
239
- async function decodeDataPageV2(cursor, header, opts) {
240
- const cursorEnd = cursor.offset + header.compressed_page_size;
241
- const valueCount = header.data_page_header_v2?.num_values;
242
- // @ts-ignore
243
- const valueCountNonNull = valueCount - header.data_page_header_v2?.num_nulls;
244
- const valueEncoding = (0, read_utils_1.getThriftEnum)(parquet_thrift_1.Encoding, header.data_page_header_v2?.encoding);
245
- /* read repetition levels */
246
- // tslint:disable-next-line:prefer-array-literal
247
- let rLevels = new Array(valueCount);
248
- if (opts.column.rLevelMax > 0) {
249
- rLevels = decodeValues(constants_1.PARQUET_RDLVL_TYPE, constants_1.PARQUET_RDLVL_ENCODING, cursor, valueCount, {
250
- bitWidth: (0, read_utils_1.getBitWidth)(opts.column.rLevelMax),
251
- disableEnvelope: true
252
- });
253
- }
254
- else {
255
- rLevels.fill(0);
256
- }
257
- /* read definition levels */
258
- // tslint:disable-next-line:prefer-array-literal
259
- let dLevels = new Array(valueCount);
260
- if (opts.column.dLevelMax > 0) {
261
- dLevels = decodeValues(constants_1.PARQUET_RDLVL_TYPE, constants_1.PARQUET_RDLVL_ENCODING, cursor, valueCount, {
262
- bitWidth: (0, read_utils_1.getBitWidth)(opts.column.dLevelMax),
263
- disableEnvelope: true
264
- });
265
- }
266
- else {
267
- dLevels.fill(0);
268
- }
269
- /* read values */
270
- let valuesBufCursor = cursor;
271
- if (header.data_page_header_v2?.is_compressed) {
272
- const valuesBuf = await (0, compression_1.decompress)(opts.compression, cursor.buffer.slice(cursor.offset, cursorEnd), header.uncompressed_page_size);
273
- valuesBufCursor = {
274
- buffer: valuesBuf,
275
- offset: 0,
276
- size: valuesBuf.length
277
- };
278
- cursor.offset = cursorEnd;
279
- }
280
- const decodeOptions = {
281
- typeLength: opts.column.typeLength,
282
- bitWidth: opts.column.typeLength
283
- };
284
- const values = decodeValues(opts.column.primitiveType, valueEncoding, valuesBufCursor, valueCountNonNull, decodeOptions);
285
- return {
286
- dlevels: dLevels,
287
- rlevels: rLevels,
288
- values,
289
- count: valueCount,
290
- pageHeader: header
291
- };
292
- }
293
- /**
294
- * Do decoding of dictionary page which helps to iterate over all indexes and get dataPage values.
295
- * @param cursor
296
- * @param pageHeader
297
- * @param options
298
- */
299
- async function decodeDictionaryPage(cursor, pageHeader, options) {
300
- const cursorEnd = cursor.offset + pageHeader.compressed_page_size;
301
- let dictCursor = {
302
- offset: 0,
303
- buffer: cursor.buffer.slice(cursor.offset, cursorEnd),
304
- size: cursorEnd - cursor.offset
305
- };
306
- cursor.offset = cursorEnd;
307
- if (options.compression !== 'UNCOMPRESSED') {
308
- const valuesBuf = await (0, compression_1.decompress)(options.compression, dictCursor.buffer.slice(dictCursor.offset, cursorEnd), pageHeader.uncompressed_page_size);
309
- dictCursor = {
310
- buffer: valuesBuf,
311
- offset: 0,
312
- size: valuesBuf.length
313
- };
314
- cursor.offset = cursorEnd;
315
- }
316
- const numValues = pageHeader?.dictionary_page_header?.num_values || 0;
317
- return decodeValues(options.column.primitiveType, options.column.encoding, dictCursor, numValues, options).map((d) => d.toString());
318
- }