Recent changes to patches

CSV : No column title when writing

2008-02-11T13:22:06Z

Add configuration method on DelimiterWriter to provide the hability to choose wether you want column title or not when writing files.

Changes on DelimiterWriter
- void setWriteColumnTitle(boolean) tell wether or not to write column title before first record.
- overwrite addRecordEntry(String columnName, Object value) to set Map capacity according to "columnTitles" size.

Also change some of AbstractWriter:
- boolean AbstractWriter.isFirstRecordWritten() that return true only after the first call to AbstractWriter.nextRecord(). Used to set column title if desired before the first record line.
- void AbstractWriter.setRowMap(Map). Used by DelimiterWriter to ajuste Map capacity when calling addRecordEntry(String , Object ).

Base on FlatPack v3.1.1

Contains unit test

Ensure unique column names

2008-02-04T09:28:02Z

I was working with flatpack using a legacy data layout with fixed column widths. To stay as close to the spec as possible, I retained the column names from the original data format specification. Unfortunately, those names were not unique (e.g. I had two columns "A1" with different lengths). Flatpack accepted my mapping file just fine and I discovered the duplicate column names only later in a debugging session.

Here's a patch to MapParser that temporarily keeps a Set of names while parsing the XML mapping and throws an IllegalArgumentException if duplicate names are detected.

Support for C style new line escaping and NULL (\N)

2006-01-27T07:36:01Z

This small patch will recognize exports that contains
at the C like newline escaping (when a newline is in a
string the exporter it will put a \ before it).
Importer must check if current line is ending with a \ and if so it must append the next line to it stripping
the \.

\N - if this sequence is found in a column (cell) it
will mark the cell as NULL.

Those two forms are used by mysql export functions
(SELECT... INTO ...) and cannot be overwriten.

Sorry for not putting a diff.

Handling fixed length files lines that are too short

2005-12-01T19:44:59Z

Paul,
I apologize for not doing as much testing on the
delimited issue as I would have liked.

I have a patch to submit to you; this enables a user to
parse files against a fixed width configuration, even
when the lines are shorter than they should be.

The reasoning behind this is I have a configuration
like this:

with lines like:
12345 2345
12345
(note no entry2 and no blank spaces for it)

Currently pzFileReader will reject the second line
because it is not long enough. In the patch I have
added a flag to allow the user to have shorter lines,
and it appends blank spaces to the end of the line so
it will parse.

I appologize that the coding isn't particularly elegant.

Currently I have only made the change for fixed width
files, but it would be useful to add it to the
delimited files also.

Luke

Add ability to cope with Blanks more correctly

2005-08-29T20:02:16Z

I have updated the DataSet.java to cope with the following:

Lines ending in a delimiter should add that column
so:
Col1,Col2,
should be interpretted with a blank Col3
this also goes for
Col1,Col2,""

Coping with blank spaces
Col1,,Col3
Should be interpretted as Blank Col2

Also
Col1,,
Should be interpretted as Blank Col2 and Col3