suddendust commented on a change in pull request #7237:
URL: https://github.com/apache/pinot/pull/7237#discussion_r690893459
##########
File path:
pinot-plugins/pinot-input-format/pinot-csv/src/main/java/org/apache/pinot/plugin/inputformat/csv/CSVRecordReader.java
##########
@@ -103,6 +105,23 @@ public void init(File dataFile, @Nullable Set<String>
fieldsToRead, @Nullable Re
_recordExtractor.init(fieldsToRead, recordExtractorConfig);
}
+ private void validateHeaderForDelimiter(char delimiter, String csvHeader,
CSVFormat format)
+ throws IOException {
+ CSVParser parser =
format.parse(RecordReaderUtils.getBufferedReader(_dataFile));
Review comment:
CSVParser doesn't move back, so once I read a record from the datafile,
it would be impossible for me to re-read it. From the docs "It is not possible
to go back, once a record has been parsed from the input stream."
##########
File path:
pinot-plugins/pinot-input-format/pinot-csv/src/main/java/org/apache/pinot/plugin/inputformat/csv/CSVRecordReader.java
##########
@@ -103,6 +105,23 @@ public void init(File dataFile, @Nullable Set<String>
fieldsToRead, @Nullable Re
_recordExtractor.init(fieldsToRead, recordExtractorConfig);
}
+ private void validateHeaderForDelimiter(char delimiter, String csvHeader,
CSVFormat format)
+ throws IOException {
+ CSVParser parser =
format.parse(RecordReaderUtils.getBufferedReader(_dataFile));
+ CSVRecord firstRecord = parser.getRecords().get(0);
+ if (recordHasMultipleValues(firstRecord) &&
delimiterNotPresentInHeader(delimiter, csvHeader)) {
Review comment:
Yes, this is much more efficient. `getRecords` is reading the entire
file. Addressed.
##########
File path:
pinot-plugins/pinot-input-format/pinot-csv/src/test/java/org/apache/pinot/plugin/inputformat/csv/CSVRecordReaderTest.java
##########
@@ -103,4 +105,74 @@ protected void checkValue(RecordReader recordReader,
List<Map<String, Object>> e
}
Assert.assertFalse(recordReader.hasNext());
}
+
+ @Test
+ public void testInvalidDelimiterInHeader() {
+ //setup
+ CSVRecordReaderConfig csvRecordReaderConfig = new CSVRecordReaderConfig();
+ csvRecordReaderConfig.setMultiValueDelimiter(CSV_MULTI_VALUE_DELIMITER);
+
csvRecordReaderConfig.setHeader("col1;col2;col3;col4;col5;col6;col7;col8;col9;col10");
+ csvRecordReaderConfig.setDelimiter(',');
+ CSVRecordReader csvRecordReader = new CSVRecordReader();
+
+ //execute and assert
+ Assert.assertThrows(IllegalArgumentException.class,
+ () -> csvRecordReader.init(_dataFile, null, csvRecordReaderConfig));
+ }
+
+ @Test
+ public void testValidDelimiterInHeader()
+ throws IOException {
+ //setup
+ CSVRecordReaderConfig csvRecordReaderConfig = new CSVRecordReaderConfig();
+ csvRecordReaderConfig.setMultiValueDelimiter(CSV_MULTI_VALUE_DELIMITER);
+
csvRecordReaderConfig.setHeader("col1,col2,col3,col4,col5,col6,col7,col8,col9,col10");
+ csvRecordReaderConfig.setDelimiter(',');
+ CSVRecordReader csvRecordReader = new CSVRecordReader();
+
+ //read all fields
+ //execute and assert
+ csvRecordReader.init(_dataFile, null, csvRecordReaderConfig);
+ Assert.assertTrue(csvRecordReader.hasNext());
+ }
+
+ /**
+ * When CSV records contain a single value, then no exception should be
throw while initialising.
+ * This test requires a different setup from the rest of the tests as it
requires a single-column
+ * CSV. Therefore, we re-write already generated records into a new file,
but only the first
+ * column.
+ *
+ * @throws IOException
+ */
+ @Test
+ public void testHeaderDelimiterSingleColumn()
+ throws IOException {
+ //setup
+
+ //create a single value CSV
+ Schema pinotSchema = getPinotSchema();
+ //write only the first column in the schema
+ String column = pinotSchema.getColumnNames().toArray(new String[0])[0];
+ //use a different file name so that other tests aren't affected
+ File file = new File(_tempDir, "data1.csv");
+ try (FileWriter fileWriter = new FileWriter(file);
+ CSVPrinter csvPrinter = new CSVPrinter(fileWriter,
CSVFormat.DEFAULT.withHeader(column))) {
+ for (Map<String, Object> r : _records) {
+ Object[] record = new Object[1];
+ for (int i = 0; i < 1; i++) {
Review comment:
Ah, I had copied this piece of code and just changed the number of
iterations from recordsMapSize to 1. Addressed.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
