[CSV-214] Adding a placeholder in the Lexer and CSV parser to store the

end-of-line string. I applied the patch in spirit and made changes:
there is no need to use a boolean to track the state of the EOL String
(set vs. not set). I also allowed for CR to be saved as an EOL string
since we allow that already.
This commit is contained in:
Gary Gregory 2017-08-11 15:05:27 -06:00
parent 4d0f226006
commit aae6f90442
4 changed files with 62 additions and 0 deletions

View File

@ -51,6 +51,7 @@
<action issue="CSV-192" type="add" dev="ggregory" due-to="Gary Gregory">Add convenience API CSVParser.parse(Path, Charset, CSVFormat)</action>
<action issue="CSV-205" type="add" dev="ggregory" due-to="Gary Gregory">Add convenience API CSVFormat#printer() to print to System.out</action>
<action issue="CSV-207" type="add" dev="ggregory" due-to="Gary Gregory">Provide a CSV Format for printing PostgreSQL CSV and Text formats.</action>
<action issue="CSV-214" type="add" dev="ggregory" due-to="Nitin Mahendru, Gary Gregory">Adding a placeholder in the Lexer and CSV parser to store the end-of-line string.</action>
</release>
<release version="1.4" date="2016-05-28" description="Feature and bug fix release">
<action issue="CSV-181" type="update" dev="ggregory" due-to="Gary Gregory">Make CSVPrinter.print(Object) GC-free.</action>

View File

@ -396,6 +396,16 @@ public final class CSVParser implements Iterable<CSVRecord>, Closeable {
return this.lexer.getCurrentLineNumber();
}
/**
* Gets the first end-of-line string encountered.
*
* @return the first end-of-line string
* @since 1.5
*/
public String getFirstEndOfLine() {
return lexer.getFirstEol();
}
/**
* Returns a copy of the header map that iterates in column order.
* <p>

View File

@ -40,6 +40,9 @@ import java.io.IOException;
*/
final class Lexer implements Closeable {
private static final String CR_STRING = Character.toString(Constants.CR);
private static final String LF_STRING = Character.toString(Constants.LF);
/**
* Constant char to use for disabling comments, escapes and encapsulation. The value -2 is used because it
* won't be confused with an EOF signal (-1), and because the Unicode value {@code FFFE} would be encoded as two
@ -57,6 +60,11 @@ final class Lexer implements Closeable {
/** The input stream */
private final ExtendedBufferedReader reader;
private String firstEol;
String getFirstEol(){
return firstEol;
}
Lexer(final CSVFormat format, final ExtendedBufferedReader reader) {
this.reader = reader;
@ -374,7 +382,20 @@ final class Lexer implements Closeable {
if (ch == CR && reader.lookAhead() == LF) {
// note: does not change ch outside of this method!
ch = reader.read();
// Save the EOL state
if (firstEol == null) {
this.firstEol = Constants.CRLF;
}
}
// save EOL state here.
if (firstEol == null) {
if (ch == LF) {
this.firstEol = LF_STRING;
} else if (ch == CR) {
this.firstEol = CR_STRING;
}
}
return ch == LF || ch == CR;
}

View File

@ -235,6 +235,36 @@ public class CSVParserTest {
}
}
@Test
public void testFirstEndOfLineCrLf() throws IOException {
final String data = "foo\r\nbaar,\r\nhello,world\r\n,kanu";
try (final CSVParser parser = CSVParser.parse(data, CSVFormat.DEFAULT)) {
final List<CSVRecord> records = parser.getRecords();
assertEquals(4, records.size());
assertEquals("\r\n", parser.getFirstEndOfLine());
}
}
@Test
public void testFirstEndOfLineLf() throws IOException {
final String data = "foo\nbaar,\nhello,world\n,kanu";
try (final CSVParser parser = CSVParser.parse(data, CSVFormat.DEFAULT)) {
final List<CSVRecord> records = parser.getRecords();
assertEquals(4, records.size());
assertEquals("\n", parser.getFirstEndOfLine());
}
}
@Test
public void testFirstEndOfLineCr() throws IOException {
final String data = "foo\rbaar,\rhello,world\r,kanu";
try (final CSVParser parser = CSVParser.parse(data, CSVFormat.DEFAULT)) {
final List<CSVRecord> records = parser.getRecords();
assertEquals(4, records.size());
assertEquals("\r", parser.getFirstEndOfLine());
}
}
@Test(expected = NoSuchElementException.class)
public void testClose() throws Exception {
final Reader in = new StringReader("# comment\na,b,c\n1,2,3\nx,y,z");