Merge pull request #295 from DamjanJovanovic/master

Add support for trailing text after the closing quote, and EOF without a final closing quote, for Excel compatibility
This commit is contained in:
Gary Gregory 2023-01-21 13:42:37 -05:00 committed by GitHub
commit b1bdb99c42
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 120 additions and 14 deletions

View File

@ -206,8 +206,12 @@ public final class CSVFormat implements Serializable {
return new Builder(csvFormat);
}
private boolean allowEofWithoutClosingQuote;
private boolean allowMissingColumnNames;
private boolean allowTrailingText;
private boolean autoFlush;
private Character commentMarker;
@ -264,6 +268,8 @@ public final class CSVFormat implements Serializable {
this.autoFlush = csvFormat.autoFlush;
this.quotedNullString = csvFormat.quotedNullString;
this.duplicateHeaderMode = csvFormat.duplicateHeaderMode;
this.allowTrailingText = csvFormat.allowTrailingText;
this.allowEofWithoutClosingQuote = csvFormat.allowEofWithoutClosingQuote;
}
/**
@ -288,6 +294,19 @@ public final class CSVFormat implements Serializable {
return this;
}
/**
* Sets whether the last field on the last line, if quoted, can have no closing quote when the file ends, {@code true} if this is ok,
* {@code false} if {@link IOException} should be thrown.
*
* @param allowEofWithoutClosingQuote whether to allow the last field on the last line to have a missing closing quote when the file ends,
* {@code true} if so, or {@code false} to cause an {@link IOException} to be thrown.
* @since 1.10.0
*/
public Builder setAllowEofWithoutClosingQuote(final boolean allowEofWithoutClosingQuote) {
this.allowEofWithoutClosingQuote = allowEofWithoutClosingQuote;
return this;
}
/**
* Sets the parser missing column names behavior, {@code true} to allow missing column names in the header line, {@code false} to cause an
* {@link IllegalArgumentException} to be thrown.
@ -301,6 +320,20 @@ public final class CSVFormat implements Serializable {
return this;
}
/**
* Sets whether to allow trailing text in a quoted field, after the closing quote.
*
* @param allowTrailingText the trailing text behavior, {@code true} to append that text to the field contents, {@code false} to throw
* an {@link IOException}.
*
* @return This instance.
* @since 1.10.0
*/
public Builder setAllowTrailingText(final boolean allowTrailingText) {
this.allowTrailingText = allowTrailingText;
return this;
}
/**
* Sets whether to flush on close.
*
@ -810,7 +843,7 @@ public final class CSVFormat implements Serializable {
* @see Predefined#Default
*/
public static final CSVFormat DEFAULT = new CSVFormat(COMMA, DOUBLE_QUOTE_CHAR, null, null, null, false, true, CRLF, null, null, null, false, false, false,
false, false, false, DuplicateHeaderMode.ALLOW_ALL);
false, false, false, DuplicateHeaderMode.ALLOW_ALL, false, false);
/**
* Excel file format (using a comma as the value delimiter). Note that the actual value delimiter used by Excel is locale dependent, it might be necessary
@ -834,6 +867,8 @@ public final class CSVFormat implements Serializable {
* <li>{@code setIgnoreEmptyLines(false)}</li>
* <li>{@code setAllowMissingColumnNames(true)}</li>
* <li>{@code setDuplicateHeaderMode(DuplicateHeaderMode.ALLOW_ALL)}</li>
* <li>{@code setAllowTrailingText(true)}</li>
* <li>{@code setAllowEofWithoutClosingQuote(true)}</li>
* </ul>
* <p>
* Note: This is currently like {@link #RFC4180} plus {@link Builder#setAllowMissingColumnNames(boolean) Builder#setAllowMissingColumnNames(true)} and
@ -846,6 +881,8 @@ public final class CSVFormat implements Serializable {
public static final CSVFormat EXCEL = DEFAULT.builder()
.setIgnoreEmptyLines(false)
.setAllowMissingColumnNames(true)
.setAllowTrailingText(true)
.setAllowEofWithoutClosingQuote(true)
.build();
// @formatter:on
@ -1268,7 +1305,7 @@ public final class CSVFormat implements Serializable {
*/
public static CSVFormat newFormat(final char delimiter) {
return new CSVFormat(String.valueOf(delimiter), null, null, null, null, false, false, null, null, null, null, false, false, false, false, false, false,
DuplicateHeaderMode.ALLOW_ALL);
DuplicateHeaderMode.ALLOW_ALL, false, false);
}
static String[] toStringArray(final Object[] values) {
@ -1310,8 +1347,12 @@ public final class CSVFormat implements Serializable {
private final DuplicateHeaderMode duplicateHeaderMode;
private final boolean allowEofWithoutClosingQuote;
private final boolean allowMissingColumnNames;
private final boolean allowTrailingText;
private final boolean autoFlush;
private final Character commentMarker; // null if commenting is disabled
@ -1366,6 +1407,8 @@ public final class CSVFormat implements Serializable {
this.autoFlush = builder.autoFlush;
this.quotedNullString = builder.quotedNullString;
this.duplicateHeaderMode = builder.duplicateHeaderMode;
this.allowTrailingText = builder.allowTrailingText;
this.allowEofWithoutClosingQuote = builder.allowEofWithoutClosingQuote;
validate();
}
@ -1396,7 +1439,7 @@ public final class CSVFormat implements Serializable {
final boolean ignoreSurroundingSpaces, final boolean ignoreEmptyLines, final String recordSeparator, final String nullString,
final Object[] headerComments, final String[] header, final boolean skipHeaderRecord, final boolean allowMissingColumnNames,
final boolean ignoreHeaderCase, final boolean trim, final boolean trailingDelimiter, final boolean autoFlush,
final DuplicateHeaderMode duplicateHeaderMode) {
final DuplicateHeaderMode duplicateHeaderMode, final boolean allowTrailingText, final boolean allowEofWithoutClosingQuote) {
this.delimiter = delimiter;
this.quoteCharacter = quoteChar;
this.quoteMode = quoteMode;
@ -1416,6 +1459,8 @@ public final class CSVFormat implements Serializable {
this.autoFlush = autoFlush;
this.quotedNullString = quoteCharacter + nullString + quoteCharacter;
this.duplicateHeaderMode = duplicateHeaderMode;
this.allowTrailingText = allowTrailingText;
this.allowEofWithoutClosingQuote = allowEofWithoutClosingQuote;
validate();
}
@ -1469,7 +1514,8 @@ public final class CSVFormat implements Serializable {
ignoreHeaderCase == other.ignoreHeaderCase && ignoreSurroundingSpaces == other.ignoreSurroundingSpaces &&
Objects.equals(nullString, other.nullString) && Objects.equals(quoteCharacter, other.quoteCharacter) && quoteMode == other.quoteMode &&
Objects.equals(quotedNullString, other.quotedNullString) && Objects.equals(recordSeparator, other.recordSeparator) &&
skipHeaderRecord == other.skipHeaderRecord && trailingDelimiter == other.trailingDelimiter && trim == other.trim;
skipHeaderRecord == other.skipHeaderRecord && trailingDelimiter == other.trailingDelimiter && trim == other.trim &&
allowTrailingText == other.allowTrailingText && allowEofWithoutClosingQuote == other.allowEofWithoutClosingQuote;
}
/**
@ -1503,6 +1549,16 @@ public final class CSVFormat implements Serializable {
return duplicateHeaderMode == DuplicateHeaderMode.ALLOW_ALL;
}
/**
* Gets whether the file can end before the last field on the last line, if quoted, has a closing quote.
*
* @return {@code true} if so, {@code false} to throw an {@link IOException}.
* @since 1.10.0
*/
public boolean getAllowEofWithoutClosingQuote() {
return allowEofWithoutClosingQuote;
}
/**
* Gets whether missing column names are allowed when parsing the header line.
*
@ -1512,6 +1568,16 @@ public final class CSVFormat implements Serializable {
return allowMissingColumnNames;
}
/**
* Gets whether quoted fields allow trailing text after the closing quote.
*
* @return {@code true} if allowed, {@code false} to throw an {@link IOException}.
* @since 1.10.0
*/
public boolean getAllowTrailingText() {
return allowTrailingText;
}
/**
* Gets whether to flush on close.
*
@ -1692,9 +1758,9 @@ public final class CSVFormat implements Serializable {
int result = 1;
result = prime * result + Arrays.hashCode(headers);
result = prime * result + Arrays.hashCode(headerComments);
return prime * result + Objects.hash(duplicateHeaderMode, allowMissingColumnNames, autoFlush, commentMarker, delimiter, escapeCharacter,
ignoreEmptyLines, ignoreHeaderCase, ignoreSurroundingSpaces, nullString, quoteCharacter, quoteMode, quotedNullString, recordSeparator,
skipHeaderRecord, trailingDelimiter, trim);
return prime * result + Objects.hash(duplicateHeaderMode, allowEofWithoutClosingQuote, allowMissingColumnNames, allowTrailingText,
autoFlush, commentMarker, delimiter, escapeCharacter, ignoreEmptyLines, ignoreHeaderCase, ignoreSurroundingSpaces,
nullString, quoteCharacter, quoteMode, quotedNullString, recordSeparator, skipHeaderRecord, trailingDelimiter, trim);
}
/**

View File

@ -57,6 +57,8 @@ final class Lexer implements Closeable {
private final boolean ignoreSurroundingSpaces;
private final boolean ignoreEmptyLines;
private final boolean allowTrailingText;
private final boolean allowEofWithoutClosingQuote;
/** The input stream */
private final ExtendedBufferedReader reader;
@ -72,6 +74,8 @@ final class Lexer implements Closeable {
this.commentStart = mapNullToDisabled(format.getCommentMarker());
this.ignoreSurroundingSpaces = format.getIgnoreSurroundingSpaces();
this.ignoreEmptyLines = format.getIgnoreEmptyLines();
this.allowTrailingText = format.getAllowTrailingText();
this.allowEofWithoutClosingQuote = format.getAllowEofWithoutClosingQuote();
this.delimiterBuf = new char[delimiter.length - 1];
this.escapeDelimiterBuf = new char[2 * delimiter.length - 1];
}
@ -364,17 +368,27 @@ final class Lexer implements Closeable {
token.type = EORECORD;
return token;
}
if (!Character.isWhitespace((char)c)) {
// error invalid char between token and next delimiter
throw new IOException("(line " + getCurrentLineNumber() +
") invalid char between encapsulated token and delimiter");
if (allowTrailingText) {
token.content.append((char) c);
} else {
if (!Character.isWhitespace((char)c)) {
// error invalid char between token and next delimiter
throw new IOException("(line " + getCurrentLineNumber() +
") invalid char between encapsulated token and delimiter");
}
}
}
}
} else if (isEndOfFile(c)) {
// error condition (end of file before end of token)
throw new IOException("(startline " + startLineNumber +
") EOF reached before encapsulated token finished");
if (allowEofWithoutClosingQuote) {
token.type = EOF;
token.isReady = true; // There is data at EOF
return token;
} else {
// error condition (end of file before end of token)
throw new IOException("(startline " + startLineNumber +
") EOF reached before encapsulated token finished");
}
} else {
// consume character
token.content.append((char) c);

View File

@ -431,4 +431,30 @@ public class LexerTest {
lexer.trimTrailingSpaces(buffer);
assertThat(lexer.nextToken(new Token()), matches(EOF, ""));
}
@Test
public void testTrailingTextAfterQuote() throws Exception {
final String code = "\"a\" b,\"a\" \" b,\"a\" b \"\"";
try (final Lexer parser = createLexer(code, CSVFormat.Builder.create().setAllowTrailingText(true).build())) {
assertThat(parser.nextToken(new Token()), matches(TOKEN, "a b"));
assertThat(parser.nextToken(new Token()), matches(TOKEN, "a \" b"));
assertThat(parser.nextToken(new Token()), matches(EOF, "a b \"\""));
}
try (final Lexer parser = createLexer(code, CSVFormat.Builder.create().setAllowTrailingText(false).build())) {
assertThrows(IOException.class, () -> parser.nextToken(new Token()));
}
}
@Test
public void testEOFWithoutClosingQuote() throws Exception {
final String code = "a,\"b";
try (final Lexer parser = createLexer(code, CSVFormat.Builder.create().setAllowEofWithoutClosingQuote(true).build())) {
assertThat(parser.nextToken(new Token()), matches(TOKEN, "a"));
assertThat(parser.nextToken(new Token()), matches(EOF, "b"));
}
try (final Lexer parser = createLexer(code, CSVFormat.Builder.create().setAllowEofWithoutClosingQuote(false).build())) {
assertThat(parser.nextToken(new Token()), matches(TOKEN, "a"));
assertThrows(IOException.class, () -> parser.nextToken(new Token()));
}
}
}