Seems useful. I had put a script for unzipping the contents of OOXML
files and expanding the XML so a human could read it somewhere in the
tree... could combine the two.
Dustin
On 1/15/16 11:10 AM, Javen O'Neal wrote:
I wrote some python code that accepts two OOXML files, unzips them
(zipfile), compares the zip structures, then compares file contents for
matching files (difflib), then outputs the diff.
Would this be helpful to the POI project, and where would I commit it?
On Jan 15, 2016 3:58 AM, "Dominik Stadler" <[email protected]> wrote:
FYI, we also have a helper application OOXMLPrettyPrint as part of the POI
source code which will take a .xlsx/.docx/.pptx/... file and produce a
pretty-printed one (which as Dave stated might not have exactly the same
content any more!). If you run this on both files you can then better
compare these. Best if you combine this with a compare tool that can be
made aware that these files are actually ZIPs and allows to do in-ZIP
comparisons...
Dominik.
On Fri, Jan 15, 2016 at 9:27 AM, David Law <[email protected]> wrote:
Please note: this may be ok for a comparison, but...
XML DOES NOT NORMALLY CONTAIN NEWLINES
They result in spurious Text nodes.
(which most products simply ignore because they are so common)
Javen O'Neals suggestion to convert to canonical is a much better idea.
Workaround: a newline BEFORE the tag-closing '>' (and '/>' ):
<html
><body bgcolor="#ffff00"
><h1>Hello world!</h1
><p
><a href="https://poi.apache.org/";>XML does not (normally)
contain newlines</a
></p
><img src="Logo.png" alt="Smiley face" border="1"
/></body
</html>
P.S. the above is valid XML but NOT Valid XHTML :-)
All the best,
DaveLaw
On 15/01/2016 04:50, Javen O'Neal wrote:
Find-replace all ">" with ">\n", then run the diff. Alternatively, use a
text editor with HTML/XML pretty printing.
On Jan 14, 2016 19:40, "Ken Hausam" <[email protected]>
wrote:
Thanks Nick. I'm a novice here. Didn't even know that .xlsx were zips of
XML files.
After changing chart type, only one file changed: the chart1.xml file
location in the xl/charts/ directory. I've attached the before
(line.xml)
and after (stackedbar.xml) files. The diff isn't enormous, but it
doesn't
seem trivial either. I couldn't list the actual diff since the XML is
all
on one line which makes the diff hard to view as a text diff.
I already tried a very simple change: changing <c:lineChart> to
<c:barChart> and changing <c:grouping val="standard"/> to <c:grouping
val="stacked"/> . No luck. Excel complained with "file format is not
valid"
when I tried to open it.
Don't know if it makes a difference, but I was using Mac Excel 2011 to
change the chart type.
Thanks,
Ken
On Wednesday, January 13, 2016 2:47 AM, Nick Burch <
[email protected]
wrote:
On Sat, 9 Jan 2016, Ken Hausam wrote:
I am using XSSFChart and associated classes to create a line chart
using
Apache POI. Works great! Thanks. My question is, is there an easy way
to
change the chart type programmatically from a line chart to a stacked
bar chart? I looked quickly at the CTChart class and associated CT
classes and noticed that the various chart types had their own class.
This makes me think that it's not as easy as just flipping a chart
type
attribute somewhere, but figured it couldn't hurt to ask.
I haven't looked at the chart stuff recently, so I can't answer off the
top of my head. What I'd suggest you do is firstly create a simple file
in
Excel, with one sheet, with a few data points, and one style of chart.
Save that. Next, change the type, and save-as that. Next, unzip both
.xlsx
files (rename to .zip and unpack). Now, compare the xml, especially for
sheets and charts, and see what differs. Post a summary of that, and
we'll
help if we can!
Nick
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
