Colin Cross | 7bb052a | 2015-02-03 12:59:37 -0800 | [diff] [blame] | 1 | // Copyright 2011 The Go Authors. All rights reserved. |
| 2 | // Use of this source code is governed by a BSD-style |
| 3 | // license that can be found in the LICENSE file. |
| 4 | |
| 5 | package csv |
| 6 | |
| 7 | import ( |
Dan Willemsen | 31b9b84 | 2021-08-31 12:51:40 -0700 | [diff] [blame] | 8 | "errors" |
| 9 | "fmt" |
Dan Willemsen | bbdf664 | 2017-01-13 22:57:23 -0800 | [diff] [blame] | 10 | "io" |
Colin Cross | 7bb052a | 2015-02-03 12:59:37 -0800 | [diff] [blame] | 11 | "reflect" |
| 12 | "strings" |
| 13 | "testing" |
Dan Willemsen | e1b3b18 | 2018-02-27 19:36:27 -0800 | [diff] [blame] | 14 | "unicode/utf8" |
Colin Cross | 7bb052a | 2015-02-03 12:59:37 -0800 | [diff] [blame] | 15 | ) |
| 16 | |
Dan Willemsen | 31b9b84 | 2021-08-31 12:51:40 -0700 | [diff] [blame] | 17 | type readTest struct { |
| 18 | Name string |
| 19 | Input string |
| 20 | Output [][]string |
| 21 | Positions [][][2]int |
| 22 | Errors []error |
Colin Cross | 7bb052a | 2015-02-03 12:59:37 -0800 | [diff] [blame] | 23 | |
Dan Willemsen | 31b9b84 | 2021-08-31 12:51:40 -0700 | [diff] [blame] | 24 | // These fields are copied into the Reader |
| 25 | Comma rune |
| 26 | Comment rune |
| 27 | UseFieldsPerRecord bool // false (default) means FieldsPerRecord is -1 |
| 28 | FieldsPerRecord int |
| 29 | LazyQuotes bool |
| 30 | TrimLeadingSpace bool |
| 31 | ReuseRecord bool |
| 32 | } |
| 33 | |
| 34 | // In these tests, the §, ¶ and ∑ characters in readTest.Input are used to denote |
| 35 | // the start of a field, a record boundary and the position of an error respectively. |
| 36 | // They are removed before parsing and are used to verify the position |
| 37 | // information reported by FieldPos. |
| 38 | |
| 39 | var readTests = []readTest{{ |
| 40 | Name: "Simple", |
| 41 | Input: "§a,§b,§c\n", |
| 42 | Output: [][]string{{"a", "b", "c"}}, |
| 43 | }, { |
| 44 | Name: "CRLF", |
| 45 | Input: "§a,§b\r\n¶§c,§d\r\n", |
| 46 | Output: [][]string{{"a", "b"}, {"c", "d"}}, |
| 47 | }, { |
| 48 | Name: "BareCR", |
| 49 | Input: "§a,§b\rc,§d\r\n", |
| 50 | Output: [][]string{{"a", "b\rc", "d"}}, |
| 51 | }, { |
| 52 | Name: "RFC4180test", |
| 53 | Input: `§#field1,§field2,§field3 |
| 54 | ¶§"aaa",§"bb |
| 55 | b",§"ccc" |
| 56 | ¶§"a,a",§"b""bb",§"ccc" |
| 57 | ¶§zzz,§yyy,§xxx |
Colin Cross | 7bb052a | 2015-02-03 12:59:37 -0800 | [diff] [blame] | 58 | `, |
Dan Willemsen | 31b9b84 | 2021-08-31 12:51:40 -0700 | [diff] [blame] | 59 | Output: [][]string{ |
| 60 | {"#field1", "field2", "field3"}, |
| 61 | {"aaa", "bb\nb", "ccc"}, |
| 62 | {"a,a", `b"bb`, "ccc"}, |
| 63 | {"zzz", "yyy", "xxx"}, |
| 64 | }, |
| 65 | UseFieldsPerRecord: true, |
| 66 | FieldsPerRecord: 0, |
| 67 | }, { |
| 68 | Name: "NoEOLTest", |
| 69 | Input: "§a,§b,§c", |
| 70 | Output: [][]string{{"a", "b", "c"}}, |
| 71 | }, { |
| 72 | Name: "Semicolon", |
| 73 | Input: "§a;§b;§c\n", |
| 74 | Output: [][]string{{"a", "b", "c"}}, |
| 75 | Comma: ';', |
| 76 | }, { |
| 77 | Name: "MultiLine", |
| 78 | Input: `§"two |
| 79 | line",§"one line",§"three |
Colin Cross | 7bb052a | 2015-02-03 12:59:37 -0800 | [diff] [blame] | 80 | line |
| 81 | field"`, |
Dan Willemsen | 31b9b84 | 2021-08-31 12:51:40 -0700 | [diff] [blame] | 82 | Output: [][]string{{"two\nline", "one line", "three\nline\nfield"}}, |
| 83 | }, { |
| 84 | Name: "BlankLine", |
| 85 | Input: "§a,§b,§c\n\n¶§d,§e,§f\n\n", |
| 86 | Output: [][]string{ |
| 87 | {"a", "b", "c"}, |
| 88 | {"d", "e", "f"}, |
| 89 | }, |
| 90 | }, { |
| 91 | Name: "BlankLineFieldCount", |
| 92 | Input: "§a,§b,§c\n\n¶§d,§e,§f\n\n", |
| 93 | Output: [][]string{ |
| 94 | {"a", "b", "c"}, |
| 95 | {"d", "e", "f"}, |
| 96 | }, |
| 97 | UseFieldsPerRecord: true, |
| 98 | FieldsPerRecord: 0, |
| 99 | }, { |
| 100 | Name: "TrimSpace", |
| 101 | Input: " §a, §b, §c\n", |
| 102 | Output: [][]string{{"a", "b", "c"}}, |
| 103 | TrimLeadingSpace: true, |
| 104 | }, { |
| 105 | Name: "LeadingSpace", |
| 106 | Input: "§ a,§ b,§ c\n", |
| 107 | Output: [][]string{{" a", " b", " c"}}, |
| 108 | }, { |
| 109 | Name: "Comment", |
| 110 | Input: "#1,2,3\n§a,§b,§c\n#comment", |
| 111 | Output: [][]string{{"a", "b", "c"}}, |
| 112 | Comment: '#', |
| 113 | }, { |
| 114 | Name: "NoComment", |
| 115 | Input: "§#1,§2,§3\n¶§a,§b,§c", |
| 116 | Output: [][]string{{"#1", "2", "3"}, {"a", "b", "c"}}, |
| 117 | }, { |
| 118 | Name: "LazyQuotes", |
| 119 | Input: `§a "word",§"1"2",§a",§"b`, |
| 120 | Output: [][]string{{`a "word"`, `1"2`, `a"`, `b`}}, |
| 121 | LazyQuotes: true, |
| 122 | }, { |
| 123 | Name: "BareQuotes", |
| 124 | Input: `§a "word",§"1"2",§a"`, |
| 125 | Output: [][]string{{`a "word"`, `1"2`, `a"`}}, |
| 126 | LazyQuotes: true, |
| 127 | }, { |
| 128 | Name: "BareDoubleQuotes", |
| 129 | Input: `§a""b,§c`, |
| 130 | Output: [][]string{{`a""b`, `c`}}, |
| 131 | LazyQuotes: true, |
| 132 | }, { |
| 133 | Name: "BadDoubleQuotes", |
| 134 | Input: `§a∑""b,c`, |
| 135 | Errors: []error{&ParseError{Err: ErrBareQuote}}, |
| 136 | }, { |
| 137 | Name: "TrimQuote", |
| 138 | Input: ` §"a",§" b",§c`, |
| 139 | Output: [][]string{{"a", " b", "c"}}, |
| 140 | TrimLeadingSpace: true, |
| 141 | }, { |
| 142 | Name: "BadBareQuote", |
| 143 | Input: `§a ∑"word","b"`, |
| 144 | Errors: []error{&ParseError{Err: ErrBareQuote}}, |
| 145 | }, { |
| 146 | Name: "BadTrailingQuote", |
| 147 | Input: `§"a word",b∑"`, |
| 148 | Errors: []error{&ParseError{Err: ErrBareQuote}}, |
| 149 | }, { |
| 150 | Name: "ExtraneousQuote", |
| 151 | Input: `§"a ∑"word","b"`, |
| 152 | Errors: []error{&ParseError{Err: ErrQuote}}, |
| 153 | }, { |
| 154 | Name: "BadFieldCount", |
| 155 | Input: "§a,§b,§c\n¶∑§d,§e", |
| 156 | Errors: []error{nil, &ParseError{Err: ErrFieldCount}}, |
| 157 | Output: [][]string{{"a", "b", "c"}, {"d", "e"}}, |
| 158 | UseFieldsPerRecord: true, |
| 159 | FieldsPerRecord: 0, |
| 160 | }, { |
| 161 | Name: "BadFieldCountMultiple", |
| 162 | Input: "§a,§b,§c\n¶∑§d,§e\n¶∑§f", |
| 163 | Errors: []error{nil, &ParseError{Err: ErrFieldCount}, &ParseError{Err: ErrFieldCount}}, |
| 164 | Output: [][]string{{"a", "b", "c"}, {"d", "e"}, {"f"}}, |
| 165 | UseFieldsPerRecord: true, |
| 166 | FieldsPerRecord: 0, |
| 167 | }, { |
| 168 | Name: "BadFieldCount1", |
| 169 | Input: `§∑a,§b,§c`, |
| 170 | Errors: []error{&ParseError{Err: ErrFieldCount}}, |
| 171 | Output: [][]string{{"a", "b", "c"}}, |
| 172 | UseFieldsPerRecord: true, |
| 173 | FieldsPerRecord: 2, |
| 174 | }, { |
| 175 | Name: "FieldCount", |
| 176 | Input: "§a,§b,§c\n¶§d,§e", |
| 177 | Output: [][]string{{"a", "b", "c"}, {"d", "e"}}, |
| 178 | }, { |
| 179 | Name: "TrailingCommaEOF", |
| 180 | Input: "§a,§b,§c,§", |
| 181 | Output: [][]string{{"a", "b", "c", ""}}, |
| 182 | }, { |
| 183 | Name: "TrailingCommaEOL", |
| 184 | Input: "§a,§b,§c,§\n", |
| 185 | Output: [][]string{{"a", "b", "c", ""}}, |
| 186 | }, { |
| 187 | Name: "TrailingCommaSpaceEOF", |
| 188 | Input: "§a,§b,§c, §", |
| 189 | Output: [][]string{{"a", "b", "c", ""}}, |
| 190 | TrimLeadingSpace: true, |
| 191 | }, { |
| 192 | Name: "TrailingCommaSpaceEOL", |
| 193 | Input: "§a,§b,§c, §\n", |
| 194 | Output: [][]string{{"a", "b", "c", ""}}, |
| 195 | TrimLeadingSpace: true, |
| 196 | }, { |
| 197 | Name: "TrailingCommaLine3", |
| 198 | Input: "§a,§b,§c\n¶§d,§e,§f\n¶§g,§hi,§", |
| 199 | Output: [][]string{{"a", "b", "c"}, {"d", "e", "f"}, {"g", "hi", ""}}, |
| 200 | TrimLeadingSpace: true, |
| 201 | }, { |
| 202 | Name: "NotTrailingComma3", |
| 203 | Input: "§a,§b,§c,§ \n", |
| 204 | Output: [][]string{{"a", "b", "c", " "}}, |
| 205 | }, { |
| 206 | Name: "CommaFieldTest", |
| 207 | Input: `§x,§y,§z,§w |
| 208 | ¶§x,§y,§z,§ |
| 209 | ¶§x,§y,§,§ |
| 210 | ¶§x,§,§,§ |
| 211 | ¶§,§,§,§ |
| 212 | ¶§"x",§"y",§"z",§"w" |
| 213 | ¶§"x",§"y",§"z",§"" |
| 214 | ¶§"x",§"y",§"",§"" |
| 215 | ¶§"x",§"",§"",§"" |
| 216 | ¶§"",§"",§"",§"" |
Colin Cross | 7bb052a | 2015-02-03 12:59:37 -0800 | [diff] [blame] | 217 | `, |
Dan Willemsen | 31b9b84 | 2021-08-31 12:51:40 -0700 | [diff] [blame] | 218 | Output: [][]string{ |
| 219 | {"x", "y", "z", "w"}, |
| 220 | {"x", "y", "z", ""}, |
| 221 | {"x", "y", "", ""}, |
| 222 | {"x", "", "", ""}, |
| 223 | {"", "", "", ""}, |
| 224 | {"x", "y", "z", "w"}, |
| 225 | {"x", "y", "z", ""}, |
| 226 | {"x", "y", "", ""}, |
| 227 | {"x", "", "", ""}, |
| 228 | {"", "", "", ""}, |
| 229 | }, |
| 230 | }, { |
| 231 | Name: "TrailingCommaIneffective1", |
| 232 | Input: "§a,§b,§\n¶§c,§d,§e", |
| 233 | Output: [][]string{ |
| 234 | {"a", "b", ""}, |
| 235 | {"c", "d", "e"}, |
| 236 | }, |
| 237 | TrimLeadingSpace: true, |
| 238 | }, { |
| 239 | Name: "ReadAllReuseRecord", |
| 240 | Input: "§a,§b\n¶§c,§d", |
| 241 | Output: [][]string{ |
| 242 | {"a", "b"}, |
| 243 | {"c", "d"}, |
| 244 | }, |
| 245 | ReuseRecord: true, |
| 246 | }, { |
| 247 | Name: "StartLine1", // Issue 19019 |
| 248 | Input: "§a,\"b\nc∑\"d,e", |
| 249 | Errors: []error{&ParseError{Err: ErrQuote}}, |
| 250 | }, { |
| 251 | Name: "StartLine2", |
| 252 | Input: "§a,§b\n¶§\"d\n\n,e∑", |
| 253 | Errors: []error{nil, &ParseError{Err: ErrQuote}}, |
| 254 | Output: [][]string{{"a", "b"}}, |
| 255 | }, { |
| 256 | Name: "CRLFInQuotedField", // Issue 21201 |
| 257 | Input: "§A,§\"Hello\r\nHi\",§B\r\n", |
| 258 | Output: [][]string{ |
| 259 | {"A", "Hello\nHi", "B"}, |
| 260 | }, |
| 261 | }, { |
| 262 | Name: "BinaryBlobField", // Issue 19410 |
| 263 | Input: "§x09\x41\xb4\x1c,§aktau", |
| 264 | Output: [][]string{{"x09A\xb4\x1c", "aktau"}}, |
| 265 | }, { |
| 266 | Name: "TrailingCR", |
| 267 | Input: "§field1,§field2\r", |
| 268 | Output: [][]string{{"field1", "field2"}}, |
| 269 | }, { |
| 270 | Name: "QuotedTrailingCR", |
| 271 | Input: "§\"field\"\r", |
| 272 | Output: [][]string{{"field"}}, |
| 273 | }, { |
| 274 | Name: "QuotedTrailingCRCR", |
| 275 | Input: "§\"field∑\"\r\r", |
| 276 | Errors: []error{&ParseError{Err: ErrQuote}}, |
| 277 | }, { |
| 278 | Name: "FieldCR", |
| 279 | Input: "§field\rfield\r", |
| 280 | Output: [][]string{{"field\rfield"}}, |
| 281 | }, { |
| 282 | Name: "FieldCRCR", |
| 283 | Input: "§field\r\rfield\r\r", |
| 284 | Output: [][]string{{"field\r\rfield\r"}}, |
| 285 | }, { |
| 286 | Name: "FieldCRCRLF", |
| 287 | Input: "§field\r\r\n¶§field\r\r\n", |
| 288 | Output: [][]string{{"field\r"}, {"field\r"}}, |
| 289 | }, { |
| 290 | Name: "FieldCRCRLFCR", |
| 291 | Input: "§field\r\r\n¶§\rfield\r\r\n\r", |
| 292 | Output: [][]string{{"field\r"}, {"\rfield\r"}}, |
| 293 | }, { |
| 294 | Name: "FieldCRCRLFCRCR", |
| 295 | Input: "§field\r\r\n¶§\r\rfield\r\r\n¶§\r\r", |
| 296 | Output: [][]string{{"field\r"}, {"\r\rfield\r"}, {"\r"}}, |
| 297 | }, { |
| 298 | Name: "MultiFieldCRCRLFCRCR", |
| 299 | Input: "§field1,§field2\r\r\n¶§\r\rfield1,§field2\r\r\n¶§\r\r,§", |
| 300 | Output: [][]string{ |
| 301 | {"field1", "field2\r"}, |
| 302 | {"\r\rfield1", "field2\r"}, |
| 303 | {"\r\r", ""}, |
| 304 | }, |
| 305 | }, { |
| 306 | Name: "NonASCIICommaAndComment", |
| 307 | Input: "§a£§b,c£ \t§d,e\n€ comment\n", |
| 308 | Output: [][]string{{"a", "b,c", "d,e"}}, |
| 309 | TrimLeadingSpace: true, |
| 310 | Comma: '£', |
| 311 | Comment: '€', |
| 312 | }, { |
| 313 | Name: "NonASCIICommaAndCommentWithQuotes", |
| 314 | Input: "§a€§\" b,\"€§ c\nλ comment\n", |
| 315 | Output: [][]string{{"a", " b,", " c"}}, |
| 316 | Comma: '€', |
| 317 | Comment: 'λ', |
| 318 | }, { |
| 319 | // λ and θ start with the same byte. |
| 320 | // This tests that the parser doesn't confuse such characters. |
| 321 | Name: "NonASCIICommaConfusion", |
| 322 | Input: "§\"abθcd\"λ§efθgh", |
| 323 | Output: [][]string{{"abθcd", "efθgh"}}, |
| 324 | Comma: 'λ', |
| 325 | Comment: '€', |
| 326 | }, { |
| 327 | Name: "NonASCIICommentConfusion", |
| 328 | Input: "§λ\n¶§λ\nθ\n¶§λ\n", |
| 329 | Output: [][]string{{"λ"}, {"λ"}, {"λ"}}, |
| 330 | Comment: 'θ', |
| 331 | }, { |
| 332 | Name: "QuotedFieldMultipleLF", |
| 333 | Input: "§\"\n\n\n\n\"", |
| 334 | Output: [][]string{{"\n\n\n\n"}}, |
| 335 | }, { |
| 336 | Name: "MultipleCRLF", |
| 337 | Input: "\r\n\r\n\r\n\r\n", |
| 338 | }, { |
| 339 | // The implementation may read each line in several chunks if it doesn't fit entirely |
| 340 | // in the read buffer, so we should test the code to handle that condition. |
| 341 | Name: "HugeLines", |
| 342 | Input: strings.Repeat("#ignore\n", 10000) + "§" + strings.Repeat("@", 5000) + ",§" + strings.Repeat("*", 5000), |
| 343 | Output: [][]string{{strings.Repeat("@", 5000), strings.Repeat("*", 5000)}}, |
| 344 | Comment: '#', |
| 345 | }, { |
| 346 | Name: "QuoteWithTrailingCRLF", |
| 347 | Input: "§\"foo∑\"bar\"\r\n", |
| 348 | Errors: []error{&ParseError{Err: ErrQuote}}, |
| 349 | }, { |
| 350 | Name: "LazyQuoteWithTrailingCRLF", |
| 351 | Input: "§\"foo\"bar\"\r\n", |
| 352 | Output: [][]string{{`foo"bar`}}, |
| 353 | LazyQuotes: true, |
| 354 | }, { |
| 355 | Name: "DoubleQuoteWithTrailingCRLF", |
| 356 | Input: "§\"foo\"\"bar\"\r\n", |
| 357 | Output: [][]string{{`foo"bar`}}, |
| 358 | }, { |
| 359 | Name: "EvenQuotes", |
| 360 | Input: `§""""""""`, |
| 361 | Output: [][]string{{`"""`}}, |
| 362 | }, { |
| 363 | Name: "OddQuotes", |
| 364 | Input: `§"""""""∑`, |
| 365 | Errors: []error{&ParseError{Err: ErrQuote}}, |
| 366 | }, { |
| 367 | Name: "LazyOddQuotes", |
| 368 | Input: `§"""""""`, |
| 369 | Output: [][]string{{`"""`}}, |
| 370 | LazyQuotes: true, |
| 371 | }, { |
| 372 | Name: "BadComma1", |
| 373 | Comma: '\n', |
| 374 | Errors: []error{errInvalidDelim}, |
| 375 | }, { |
| 376 | Name: "BadComma2", |
| 377 | Comma: '\r', |
| 378 | Errors: []error{errInvalidDelim}, |
| 379 | }, { |
| 380 | Name: "BadComma3", |
| 381 | Comma: '"', |
| 382 | Errors: []error{errInvalidDelim}, |
| 383 | }, { |
| 384 | Name: "BadComma4", |
| 385 | Comma: utf8.RuneError, |
| 386 | Errors: []error{errInvalidDelim}, |
| 387 | }, { |
| 388 | Name: "BadComment1", |
| 389 | Comment: '\n', |
| 390 | Errors: []error{errInvalidDelim}, |
| 391 | }, { |
| 392 | Name: "BadComment2", |
| 393 | Comment: '\r', |
| 394 | Errors: []error{errInvalidDelim}, |
| 395 | }, { |
| 396 | Name: "BadComment3", |
| 397 | Comment: utf8.RuneError, |
| 398 | Errors: []error{errInvalidDelim}, |
| 399 | }, { |
| 400 | Name: "BadCommaComment", |
| 401 | Comma: 'X', |
| 402 | Comment: 'X', |
| 403 | Errors: []error{errInvalidDelim}, |
| 404 | }} |
Colin Cross | 7bb052a | 2015-02-03 12:59:37 -0800 | [diff] [blame] | 405 | |
Dan Willemsen | 31b9b84 | 2021-08-31 12:51:40 -0700 | [diff] [blame] | 406 | func TestRead(t *testing.T) { |
| 407 | newReader := func(tt readTest) (*Reader, [][][2]int, map[int][2]int) { |
| 408 | positions, errPositions, input := makePositions(tt.Input) |
| 409 | r := NewReader(strings.NewReader(input)) |
| 410 | |
| 411 | if tt.Comma != 0 { |
| 412 | r.Comma = tt.Comma |
| 413 | } |
| 414 | r.Comment = tt.Comment |
| 415 | if tt.UseFieldsPerRecord { |
| 416 | r.FieldsPerRecord = tt.FieldsPerRecord |
| 417 | } else { |
| 418 | r.FieldsPerRecord = -1 |
| 419 | } |
| 420 | r.LazyQuotes = tt.LazyQuotes |
| 421 | r.TrimLeadingSpace = tt.TrimLeadingSpace |
| 422 | r.ReuseRecord = tt.ReuseRecord |
| 423 | return r, positions, errPositions |
| 424 | } |
| 425 | |
| 426 | for _, tt := range readTests { |
Dan Willemsen | e1b3b18 | 2018-02-27 19:36:27 -0800 | [diff] [blame] | 427 | t.Run(tt.Name, func(t *testing.T) { |
Dan Willemsen | 31b9b84 | 2021-08-31 12:51:40 -0700 | [diff] [blame] | 428 | r, positions, errPositions := newReader(tt) |
Dan Willemsen | e1b3b18 | 2018-02-27 19:36:27 -0800 | [diff] [blame] | 429 | out, err := r.ReadAll() |
Dan Willemsen | 31b9b84 | 2021-08-31 12:51:40 -0700 | [diff] [blame] | 430 | if wantErr := firstError(tt.Errors, positions, errPositions); wantErr != nil { |
| 431 | if !reflect.DeepEqual(err, wantErr) { |
| 432 | t.Fatalf("ReadAll() error mismatch:\ngot %v (%#v)\nwant %v (%#v)", err, err, wantErr, wantErr) |
| 433 | } |
| 434 | if out != nil { |
| 435 | t.Fatalf("ReadAll() output:\ngot %q\nwant nil", out) |
| 436 | } |
| 437 | } else { |
| 438 | if err != nil { |
| 439 | t.Fatalf("unexpected Readall() error: %v", err) |
| 440 | } |
| 441 | if !reflect.DeepEqual(out, tt.Output) { |
| 442 | t.Fatalf("ReadAll() output:\ngot %q\nwant %q", out, tt.Output) |
| 443 | } |
| 444 | } |
| 445 | |
| 446 | // Check field and error positions. |
| 447 | r, _, _ = newReader(tt) |
| 448 | for recNum := 0; ; recNum++ { |
| 449 | rec, err := r.Read() |
| 450 | var wantErr error |
| 451 | if recNum < len(tt.Errors) && tt.Errors[recNum] != nil { |
| 452 | wantErr = errorWithPosition(tt.Errors[recNum], recNum, positions, errPositions) |
| 453 | } else if recNum >= len(tt.Output) { |
| 454 | wantErr = io.EOF |
| 455 | } |
| 456 | if !reflect.DeepEqual(err, wantErr) { |
| 457 | t.Fatalf("Read() error at record %d:\ngot %v (%#v)\nwant %v (%#v)", recNum, err, err, wantErr, wantErr) |
| 458 | } |
| 459 | // ErrFieldCount is explicitly non-fatal. |
| 460 | if err != nil && !errors.Is(err, ErrFieldCount) { |
| 461 | if recNum < len(tt.Output) { |
| 462 | t.Fatalf("need more records; got %d want %d", recNum, len(tt.Output)) |
| 463 | } |
| 464 | break |
| 465 | } |
| 466 | if got, want := rec, tt.Output[recNum]; !reflect.DeepEqual(got, want) { |
| 467 | t.Errorf("Read vs ReadAll mismatch;\ngot %q\nwant %q", got, want) |
| 468 | } |
| 469 | pos := positions[recNum] |
| 470 | if len(pos) != len(rec) { |
| 471 | t.Fatalf("mismatched position length at record %d", recNum) |
| 472 | } |
| 473 | for i := range rec { |
| 474 | line, col := r.FieldPos(i) |
| 475 | if got, want := [2]int{line, col}, pos[i]; got != want { |
| 476 | t.Errorf("position mismatch at record %d, field %d;\ngot %v\nwant %v", recNum, i, got, want) |
| 477 | } |
| 478 | } |
Dan Willemsen | e1b3b18 | 2018-02-27 19:36:27 -0800 | [diff] [blame] | 479 | } |
| 480 | }) |
Colin Cross | 7bb052a | 2015-02-03 12:59:37 -0800 | [diff] [blame] | 481 | } |
| 482 | } |
Dan Willemsen | 6ff2325 | 2015-09-15 13:49:18 -0700 | [diff] [blame] | 483 | |
Dan Willemsen | 31b9b84 | 2021-08-31 12:51:40 -0700 | [diff] [blame] | 484 | // firstError returns the first non-nil error in errs, |
| 485 | // with the position adjusted according to the error's |
| 486 | // index inside positions. |
| 487 | func firstError(errs []error, positions [][][2]int, errPositions map[int][2]int) error { |
| 488 | for i, err := range errs { |
| 489 | if err != nil { |
| 490 | return errorWithPosition(err, i, positions, errPositions) |
| 491 | } |
| 492 | } |
| 493 | return nil |
| 494 | } |
| 495 | |
| 496 | func errorWithPosition(err error, recNum int, positions [][][2]int, errPositions map[int][2]int) error { |
| 497 | parseErr, ok := err.(*ParseError) |
| 498 | if !ok { |
| 499 | return err |
| 500 | } |
| 501 | if recNum >= len(positions) { |
| 502 | panic(fmt.Errorf("no positions found for error at record %d", recNum)) |
| 503 | } |
| 504 | errPos, ok := errPositions[recNum] |
| 505 | if !ok { |
| 506 | panic(fmt.Errorf("no error position found for error at record %d", recNum)) |
| 507 | } |
| 508 | parseErr1 := *parseErr |
| 509 | parseErr1.StartLine = positions[recNum][0][0] |
| 510 | parseErr1.Line = errPos[0] |
| 511 | parseErr1.Column = errPos[1] |
| 512 | return &parseErr1 |
| 513 | } |
| 514 | |
| 515 | // makePositions returns the expected field positions of all |
| 516 | // the fields in text, the positions of any errors, and the text with the position markers |
| 517 | // removed. |
| 518 | // |
| 519 | // The start of each field is marked with a § symbol; |
| 520 | // CSV lines are separated by ¶ symbols; |
| 521 | // Error positions are marked with ∑ symbols. |
| 522 | func makePositions(text string) ([][][2]int, map[int][2]int, string) { |
| 523 | buf := make([]byte, 0, len(text)) |
| 524 | var positions [][][2]int |
| 525 | errPositions := make(map[int][2]int) |
| 526 | line, col := 1, 1 |
| 527 | recNum := 0 |
| 528 | |
| 529 | for len(text) > 0 { |
| 530 | r, size := utf8.DecodeRuneInString(text) |
| 531 | switch r { |
| 532 | case '\n': |
| 533 | line++ |
| 534 | col = 1 |
| 535 | buf = append(buf, '\n') |
| 536 | case '§': |
| 537 | if len(positions) == 0 { |
| 538 | positions = append(positions, [][2]int{}) |
| 539 | } |
| 540 | positions[len(positions)-1] = append(positions[len(positions)-1], [2]int{line, col}) |
| 541 | case '¶': |
| 542 | positions = append(positions, [][2]int{}) |
| 543 | recNum++ |
| 544 | case '∑': |
| 545 | errPositions[recNum] = [2]int{line, col} |
| 546 | default: |
| 547 | buf = append(buf, text[:size]...) |
| 548 | col += size |
| 549 | } |
| 550 | text = text[size:] |
| 551 | } |
| 552 | return positions, errPositions, string(buf) |
| 553 | } |
| 554 | |
Dan Willemsen | bbdf664 | 2017-01-13 22:57:23 -0800 | [diff] [blame] | 555 | // nTimes is an io.Reader which yields the string s n times. |
| 556 | type nTimes struct { |
| 557 | s string |
| 558 | n int |
| 559 | off int |
| 560 | } |
| 561 | |
| 562 | func (r *nTimes) Read(p []byte) (n int, err error) { |
| 563 | for { |
| 564 | if r.n <= 0 || r.s == "" { |
| 565 | return n, io.EOF |
| 566 | } |
| 567 | n0 := copy(p, r.s[r.off:]) |
| 568 | p = p[n0:] |
| 569 | n += n0 |
| 570 | r.off += n0 |
| 571 | if r.off == len(r.s) { |
| 572 | r.off = 0 |
| 573 | r.n-- |
| 574 | } |
| 575 | if len(p) == 0 { |
| 576 | return |
| 577 | } |
| 578 | } |
| 579 | } |
| 580 | |
| 581 | // benchmarkRead measures reading the provided CSV rows data. |
| 582 | // initReader, if non-nil, modifies the Reader before it's used. |
| 583 | func benchmarkRead(b *testing.B, initReader func(*Reader), rows string) { |
| 584 | b.ReportAllocs() |
| 585 | r := NewReader(&nTimes{s: rows, n: b.N}) |
| 586 | if initReader != nil { |
| 587 | initReader(r) |
| 588 | } |
| 589 | for { |
| 590 | _, err := r.Read() |
| 591 | if err == io.EOF { |
| 592 | break |
| 593 | } |
| 594 | if err != nil { |
| 595 | b.Fatal(err) |
| 596 | } |
| 597 | } |
| 598 | } |
| 599 | |
| 600 | const benchmarkCSVData = `x,y,z,w |
Dan Willemsen | 6ff2325 | 2015-09-15 13:49:18 -0700 | [diff] [blame] | 601 | x,y,z, |
| 602 | x,y,, |
| 603 | x,,, |
| 604 | ,,, |
| 605 | "x","y","z","w" |
| 606 | "x","y","z","" |
| 607 | "x","y","","" |
| 608 | "x","","","" |
| 609 | "","","","" |
| 610 | ` |
| 611 | |
Dan Willemsen | bbdf664 | 2017-01-13 22:57:23 -0800 | [diff] [blame] | 612 | func BenchmarkRead(b *testing.B) { |
| 613 | benchmarkRead(b, nil, benchmarkCSVData) |
| 614 | } |
Dan Willemsen | 6ff2325 | 2015-09-15 13:49:18 -0700 | [diff] [blame] | 615 | |
Dan Willemsen | bbdf664 | 2017-01-13 22:57:23 -0800 | [diff] [blame] | 616 | func BenchmarkReadWithFieldsPerRecord(b *testing.B) { |
| 617 | benchmarkRead(b, func(r *Reader) { r.FieldsPerRecord = 4 }, benchmarkCSVData) |
| 618 | } |
| 619 | |
| 620 | func BenchmarkReadWithoutFieldsPerRecord(b *testing.B) { |
| 621 | benchmarkRead(b, func(r *Reader) { r.FieldsPerRecord = -1 }, benchmarkCSVData) |
| 622 | } |
| 623 | |
| 624 | func BenchmarkReadLargeFields(b *testing.B) { |
| 625 | benchmarkRead(b, nil, strings.Repeat(`xxxxxxxxxxxxxxxx,yyyyyyyyyyyyyyyy,zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz,wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww,vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv |
| 626 | xxxxxxxxxxxxxxxxxxxxxxxx,yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy,zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz,wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww,vvvv |
| 627 | ,,zzzz,wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww,vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv |
| 628 | xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx,yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy,zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz,wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww,vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv |
| 629 | `, 3)) |
Dan Willemsen | 6ff2325 | 2015-09-15 13:49:18 -0700 | [diff] [blame] | 630 | } |
Dan Willemsen | c78f714 | 2017-07-26 13:08:14 -0700 | [diff] [blame] | 631 | |
| 632 | func BenchmarkReadReuseRecord(b *testing.B) { |
| 633 | benchmarkRead(b, func(r *Reader) { r.ReuseRecord = true }, benchmarkCSVData) |
| 634 | } |
| 635 | |
| 636 | func BenchmarkReadReuseRecordWithFieldsPerRecord(b *testing.B) { |
| 637 | benchmarkRead(b, func(r *Reader) { r.ReuseRecord = true; r.FieldsPerRecord = 4 }, benchmarkCSVData) |
| 638 | } |
| 639 | |
| 640 | func BenchmarkReadReuseRecordWithoutFieldsPerRecord(b *testing.B) { |
| 641 | benchmarkRead(b, func(r *Reader) { r.ReuseRecord = true; r.FieldsPerRecord = -1 }, benchmarkCSVData) |
| 642 | } |
| 643 | |
| 644 | func BenchmarkReadReuseRecordLargeFields(b *testing.B) { |
| 645 | benchmarkRead(b, func(r *Reader) { r.ReuseRecord = true }, strings.Repeat(`xxxxxxxxxxxxxxxx,yyyyyyyyyyyyyyyy,zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz,wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww,vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv |
| 646 | xxxxxxxxxxxxxxxxxxxxxxxx,yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy,zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz,wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww,vvvv |
| 647 | ,,zzzz,wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww,vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv |
| 648 | xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx,yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy,zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz,wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww,vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv |
| 649 | `, 3)) |
| 650 | } |