Examples with LineDecoder org.embulk.spi.util.LineDecoder used on opensource projects

Example 1 with LineDecoder

use of org.embulk.spi.util.LineDecoder in project embulk by embulk.

the class CsvParserPlugin method run.

@Override
public void run(TaskSource taskSource, final Schema schema, FileInput input, PageOutput output) {
    PluginTask task = taskSource.loadTask(PluginTask.class);
    final TimestampParser[] timestampParsers = Timestamps.newTimestampColumnParsers(task, task.getSchemaConfig());
    final JsonParser jsonParser = new JsonParser();
    final CsvTokenizer tokenizer = new CsvTokenizer(new LineDecoder(input, task), task);
    final boolean allowOptionalColumns = task.getAllowOptionalColumns();
    final boolean allowExtraColumns = task.getAllowExtraColumns();
    final boolean stopOnInvalidRecord = task.getStopOnInvalidRecord();
    final int skipHeaderLines = task.getSkipHeaderLines();
    try (final PageBuilder pageBuilder = new PageBuilder(Exec.getBufferAllocator(), schema, output)) {
        while (tokenizer.nextFile()) {
            // skip the header lines for each file
            for (int skipHeaderLineNumber = skipHeaderLines; skipHeaderLineNumber > 0; skipHeaderLineNumber--) {
                if (!tokenizer.skipHeaderLine()) {
                    break;
                }
            }
            if (!tokenizer.nextRecord()) {
                // empty file
                continue;
            }
            while (true) {
                boolean hasNextRecord;
                try {
                    schema.visitColumns(new ColumnVisitor() {

                        public void booleanColumn(Column column) {
                            String v = nextColumn();
                            if (v == null) {
                                pageBuilder.setNull(column);
                            } else {
                                pageBuilder.setBoolean(column, TRUE_STRINGS.contains(v));
                            }
                        }

                        public void longColumn(Column column) {
                            String v = nextColumn();
                            if (v == null) {
                                pageBuilder.setNull(column);
                            } else {
                                try {
                                    pageBuilder.setLong(column, Long.parseLong(v));
                                } catch (NumberFormatException e) {
                                    // TODO support default value
                                    throw new CsvRecordValidateException(e);
                                }
                            }
                        }

                        public void doubleColumn(Column column) {
                            String v = nextColumn();
                            if (v == null) {
                                pageBuilder.setNull(column);
                            } else {
                                try {
                                    pageBuilder.setDouble(column, Double.parseDouble(v));
                                } catch (NumberFormatException e) {
                                    // TODO support default value
                                    throw new CsvRecordValidateException(e);
                                }
                            }
                        }

                        public void stringColumn(Column column) {
                            String v = nextColumn();
                            if (v == null) {
                                pageBuilder.setNull(column);
                            } else {
                                pageBuilder.setString(column, v);
                            }
                        }

                        public void timestampColumn(Column column) {
                            String v = nextColumn();
                            if (v == null) {
                                pageBuilder.setNull(column);
                            } else {
                                try {
                                    pageBuilder.setTimestamp(column, timestampParsers[column.getIndex()].parse(v));
                                } catch (TimestampParseException e) {
                                    // TODO support default value
                                    throw new CsvRecordValidateException(e);
                                }
                            }
                        }

                        public void jsonColumn(Column column) {
                            String v = nextColumn();
                            if (v == null) {
                                pageBuilder.setNull(column);
                            } else {
                                try {
                                    pageBuilder.setJson(column, jsonParser.parse(v));
                                } catch (JsonParseException e) {
                                    // TODO support default value
                                    throw new CsvRecordValidateException(e);
                                }
                            }
                        }

                        private String nextColumn() {
                            if (allowOptionalColumns && !tokenizer.hasNextColumn()) {
                                // TODO warning
                                return null;
                            }
                            return tokenizer.nextColumnOrNull();
                        }
                    });
                    try {
                        hasNextRecord = tokenizer.nextRecord();
                    } catch (CsvTokenizer.TooManyColumnsException ex) {
                        if (allowExtraColumns) {
                            String tooManyColumnsLine = tokenizer.skipCurrentLine();
                            // TODO warning
                            hasNextRecord = tokenizer.nextRecord();
                        } else {
                            // this line will be skipped at the following catch section
                            throw ex;
                        }
                    }
                    pageBuilder.addRecord();
                } catch (CsvTokenizer.InvalidFormatException | CsvTokenizer.InvalidValueException | CsvRecordValidateException e) {
                    String skippedLine = tokenizer.skipCurrentLine();
                    long lineNumber = tokenizer.getCurrentLineNumber();
                    if (stopOnInvalidRecord) {
                        throw new DataException(String.format("Invalid record at line %d: %s", lineNumber, skippedLine), e);
                    }
                    log.warn(String.format("Skipped line %d (%s): %s", lineNumber, e.getMessage(), skippedLine));
                    // exec.notice().skippedLine(skippedLine);
                    hasNextRecord = tokenizer.nextRecord();
                }
                if (!hasNextRecord) {
                    break;
                }
            }
        }
        pageBuilder.finish();
    }
}

Example 2 with LineDecoder

use of org.embulk.spi.util.LineDecoder in project embulk by embulk.

the class TestCsvTokenizer method recoverFromQuotedSizeLimitExceededException.

@Test
public void recoverFromQuotedSizeLimitExceededException() throws Exception {
    config.set("max_quoted_size_limit", 12);
    reloadPluginTask();
    String[] lines = new String[] { "v1,v2", // this is a broken line and should be skipped
    "v3,\"0123", // this line should be not be skiped
    "v4,v5", // this line should be not be skiped
    "v6,v7" };
    FileInput input = newFileInputFromLines(task, lines);
    LineDecoder decoder = new LineDecoder(input, task);
    CsvTokenizer tokenizer = new CsvTokenizer(decoder, task);
    Schema schema = task.getSchemaConfig().toSchema();
    tokenizer.nextFile();
    assertTrue(tokenizer.nextRecord());
    assertEquals("v1", tokenizer.nextColumn());
    assertEquals("v2", tokenizer.nextColumn());
    assertTrue(tokenizer.nextRecord());
    assertEquals("v3", tokenizer.nextColumn());
    try {
        tokenizer.nextColumn();
        fail();
    } catch (Exception e) {
        assertTrue(e instanceof CsvTokenizer.QuotedSizeLimitExceededException);
    }
    assertEquals("v3,\"0123", tokenizer.skipCurrentLine());
    assertTrue(tokenizer.nextRecord());
    assertEquals("v4", tokenizer.nextColumn());
    assertEquals("v5", tokenizer.nextColumn());
    assertTrue(tokenizer.nextRecord());
    assertEquals("v6", tokenizer.nextColumn());
    assertEquals("v7", tokenizer.nextColumn());
}

Example 3 with LineDecoder

use of org.embulk.spi.util.LineDecoder in project embulk by embulk.

the class TestCsvTokenizer method parse.

private static List<List<String>> parse(CsvParserPlugin.PluginTask task, FileInput input) {
    LineDecoder decoder = new LineDecoder(input, task);
    CsvTokenizer tokenizer = new CsvTokenizer(decoder, task);
    Schema schema = task.getSchemaConfig().toSchema();
    tokenizer.nextFile();
    List<List<String>> records = new ArrayList<>();
    while (tokenizer.nextRecord()) {
        List<String> record = new ArrayList<>();
        for (Column c : schema.getColumns()) {
            String v = tokenizer.nextColumnOrNull();
            record.add(v);
        }
        records.add(record);
    }
    return records;
}