/**
* Parses a CSV header line and creates the corresponding resulting schema in the section (i.e.: a
* set of fields with positions)
*/
protected List<Field> parseHeaders(List<String[]> csvLines) {
Section section = sectionInstance.getSection();
String sectionId = section.getId();
String headers[] = csvLines.get(0);
log.trace("Headers for section " + sectionId + ": " + Arrays.toString(headers));
int nlines = csvLines.size();
for (int j = 0; j < headers.length; j++) {
String header = StringUtils.trimToNull(headers[j]);
if (header == null) {
// Empty header, let's remove all!
//
log.warn(i18n.msg("empty_column", j));
for (int i = 0; i < nlines; i++) {
String[] line = csvLines.get(i);
if (j == line.length)
// This happens sometime: the header has one extra-tab and the lines are OK
// we need to recraft the header only in that case
{
continue;
}
csvLines.set(i, (String[]) ArrayUtils.remove(line, j));
}
headers = csvLines.get(0);
j--; // Restart from this point, which is now the next column
continue;
}
log.trace("Working on header " + j + ":" + header);
Field field = section.getFieldByHeader(header, false);
if (field == null) {
throw new TabValidationException(
i18n.msg("unexpected_field_in_section_error", header, sectionId));
}
// Let's add a new real field on the basis of the header. The new field created gets its id
// from the
// original one in the schema, so it will have the "canonical" form, independently on what we
// found
// on the input (e.g.: upper case).
//
field = field.parseHeader(header, j, false);
sectionInstance.addField(field);
}
return sectionInstance.getFields();
}
public SectionInstance parseCsvLines(
List<String[]> csvLines, int fromRow, int toRow, int fromCol, int toCol) {
Section section = sectionInstance.getSection();
String sectionId = section.getId();
String fileId = sectionInstance.getParent().getFileId();
log.trace(
String.format(
"___ Parsing 'lines' of Section: %s ('%s', %d, %d, %d, %d) ___\n",
sectionId, fileId, fromRow, toRow, fromCol, toCol));
if (csvLines == null || csvLines.size() == 0) {
log.trace("WARNING, The CSV lines are empty, section: " + sectionId);
return sectionInstance;
}
if (toCol == -1) {
toCol = csvLines.get(0).length - 1;
}
if (fromRow == 0) {
fromRow = 1;
}
if (toRow - fromRow < 0 || toCol - fromCol < 0) {
log.trace("WARNING, The specified window is empty, section " + sectionId);
return sectionInstance;
}
List<Field> fields = parseHeaders(csvLines);
toCol = Math.min(toCol, fields.size() - 1);
// Let's go over the lines
//
for (int i = fromRow; i <= toRow; i++) {
String[] line = csvLines.get(i);
if (line == null) {
log.debug("Line #" + i + " is null, skipping");
continue;
}
// Let's go along the headers
//
Record record = new Record(sectionInstance);
for (int icol = fromCol; icol <= toCol; icol++) {
// Set up the current record's value
//
Field field = sectionInstance.getField(icol);
if (field == null) {
throw new TabInternalErrorException(
i18n.msg("generic_field_error", icol, section.getId(), i));
}
if (icol >= line.length) {
log.debug("The (logical) line #" + i + " is too short, ignoring all from col #" + icol);
break;
}
String value = line[icol];
record.set(icol, value);
}
// Let's add the record, provided that it is not empty
if (record.isNull()) {
log.trace("Parsing records for section " + sectionId + ": empty record");
} else {
if (log.isTraceEnabled()) {
log.trace("adding the record:" + record + " to the section " + sectionId);
}
sectionInstance.addRecord(record);
}
} // loop on the lines
log.trace("___ /end of parsing section " + sectionId + " ___");
return sectionInstance;
}