Author: Kohei Yoshida

LibreOffice contributor, spreadsheet nerd, and machine learning beginner.

Custom sort in DataPilot

Kohei Yoshida

Just checked in this piece to the master branch of the Go-OO repository, to support sorting of DataPilot’s field members using custom sort lists. I’ve extended the popup window I wrote for the hide field members functionality to provide this additional sorting functionality. The result is the following popup window:

In the upper half of the window I’ve added a menu-like control, with the custom sort lists being provided in the submenu. The UI is fully functional, but still a bit rough around the edges. The custom sort list submenu, in particular, may need some additional work to handle a large set of custom sort lists, a sort list that is very long, or stuff like that. But as long as your sort list is in modest size, it should work just fine.

This feature didn’t make it in to 3.1 since we are in a stabilization phase for 3.1. But as soon as we branch master for the stable 3.1 branch, I will enable this feature in the default build in the master branch.

ODF Support in MS Office 2007

Kohei Yoshida

I guess it’s all over the news right now, that the latest service pack (SP2) for MS Office 2007 will enable Office to import and export ODF natively. This blog piece by Doug Mahugh touches on the word processor part of their ODF support. I haven’t yet tried it myself, but judging by Doug’s blog article it looks pretty impressive.

But, being more of a spreadsheet person, I’m personally more interested in how it fares in Excel to Calc interoperability. Since their ODF support is on ODF 1.1, which predates the on-going OpenFormula specification work, I’d be interested to see how compatible the formulas are. Technically speaking, as of ODF 1.1, interpreting formula expressions was pretty much application-specific, so I would not be surprised even if they are not compatible at all. But we’ll just have to see.

Either way, I find this news very encouraging. This is undoubtedly a big step toward proliferation of ODF as a practical document exchange format.

Traversing to precedents and dependents

Kohei Yoshida

Just checked this piece into the master branch of the go-oo repository. It allows traversing to precedents and dependents of a cell by Ctrl-[ and Ctrl-] key strokes, respectively. It is similar in concept to the existing Detective functionality, but while Detective graphically shows the precedents and dependents of a current cell, this new functionality physically moves the cursor to the precedent and dependent cells. Similar functionality already existed in Excel, so this is one of those interoperability features and, for some spreadsheet users, the ability to jump to precedents/dependents is apparently very important for their productivity.

No feature is complete without screenshots. So, here they are. :-)

In the first example, the cell currently selected contains references to three cells and one cell range in its formula expression. I’m showing the precedent traces here just to show their relationship visually.

When you hit Ctrl-[, it highlights all its reference cells (a.k.a precedents) and moves the cursor to the first precedent.

Once the cells are highlighted, you can easily navigate through the highlighted cells by hitting the Enter or Tab key.

There is one caveat. When the expression includes references to cells outside of the current sheet, the ones that are not on the current sheet are ignored. The only exception to this rule is when the first reference points to a cell in another sheet, in which case it jumps to that external-sheet cell while the rest of the references are ignored even if they are on the current sheet. If that reference is in another document (i.e. external reference), it opens that document then sets the cursor to the referenced cell position provided that the document is available at specified location in the file system.

Traversing to dependents also works in a similar fashion. Consider the following example

where current cell is referenced by multiple other cells. Again, I’m showing the dependency traces to display their relationship graphically. When you hit Ctrl-], it highlights all its dependent cells and moves the cursor to the first dependent cell.

Similar to the precedent jump, there is a caveat; when the cell is referenced by cells on multiple different sheets including the current sheet, only those cells on the current sheet are highlighted and the rest are ignored. But unlike the precedent jump, there is no exception to this rule.

That’s it folks! I hope you find this new functionality useful. And as always, please report back any problems you may encounter so that I can fix them. Thank you very much, ladies and gentlemen. :-)

Some minor enhancement with DataPilot

Kohei Yoshida2 Comments

Just checked into the master branch of go-oo repository (oh BTW we just switched our repository from svn on gnome.org to git on freedesktop.org) is a minor enhancement in DataPilot, to allow the users to filter results by field members directly from the field buttons in the table output. The following screenshot will tell you the change I just made:

The new field buttons also provide visual feedback on fields that are filtered; the same way the autofilter buttons already do. This way it’s visually obvious to the users which fields are currently filtered.

I call this a “minor” enhancement, simply because it doesn’t provide a new functionality per se; the same filtering functionality was already present but hidden deep beneath multiple layers of dialogs. You needed to go from the datapilot layout dialog (the main dialog), open up the field dialog for the field you want to filter results by, and then finally the field options dialog to get to the Hide items list in that dialog. This enhancement will push that functionality up-front and make it more accessible to the users, along with some visual feedback.

Anyway, this one made it into 3.1, so testing and feedback is greatly appreciated. :-)

Coloring tabs in Calc

Kohei Yoshida7 Comments

I just checked in to go-oo trunk the latest patch from Daniel Watson that enables coloring of sheet tabs in Calc. With his latest patch, the tab color dialog looks more compact and polished, which I definitely prefer over the previous version.

In case it’s not obvious how to use this feature, here are some screenshots to show how.

In the context menu on the sheet tab, there is now a new entry called Tab Color.

Select that menu to bring up the color palette dialog.

Select a color and click OK. The sheet tab now has a sliver of color at the bottom. When you switch to another sheet, you’ll notice that the tab you just colored is now filled entirely with that color. And when you color all of your tabs, you’ll get something like this:

Pretty cool huh?

BTW Daniel did this almost entirely on his own, with just a teeny bit of help from Eric Bachard and myself. He is very thorough and the quality of his code is pretty high. He maintains the spec for this feature right here. He is also anxious to get this feature upstreamed, and judging by the kind of reaction he is getting from the upstream Calc team, I’m pretty sure this feature is on a fast track to upstream acceptance. ;-) Great job Daniel!

Oh, one last thing. Please try out this feature and test it out.

It’s text when it’s quoted, stupid!

Kohei Yoshida4 Comments

Ok. This is another one I just checked into ooo-build trunk. It’s actually a very minor enhancement, but some people may find this useful.

What I did is to modify Calc’s csv import filter to provide the users an option to always import quoted fields (or cells) as text. Here is a screenshot of the dialog with this new option:

Hopefully what it does is self-evident. If not, here is an example.

Let’s say you have a csv file of the following content:

"0-0" "1-0" "2-0" "3-0" "4-0" "5-0"
"0-1" "1-1" "2-1" "3-1" "4-1" "5-1"

When you import this file into Calc, here is what you get with the current version:

As you can see, most of the fields got converted to dates. You get the same results even if those individual cells are not quoted, because Calc’s current csv import filter doesn’t make any distinction between quoted and unquoted fields in terms of data types.

When you import this file using the new option, you get this instead:

Now, all quoted fields are imported as text, with no surprise conversion. Plain and simple. :-)

ike-scan available from Build Service (and how I built my first RPM package)

Kohei Yoshida2 Comments

ike-scan is now available for openSUSE from my personal repository on openSUSE Build Service. I decided to provide a package for this network security tool because my wife needed it for her job, and had asked me why no one had provided an RPM package for this tool on openSUSE.

This was also the first time I ever built any RPM package on my own. There was a lot of tutorials all over the web on how to build RPM packages, but many were either very outdated or didn’t give me enough info to fully understand the RPM packaging process (or I was just having a bad Google karma). I ended up wasting more time than I had wanted out of my precious Saturday family time, but eventually got the package to build. (Yay for me!)

What’s more cool, though, is the Build Service that openSUSE project provides. The Build Service basically provides a convenient way to build and host RPM packages for multiple distros, not just openSUSE but other distros like Fedora, RHEL, Debian and Ubuntu, while all you have to do is to provide the source tarball and a spec file. There is even a wizard to create a spec file from a tarball (though you probably have to review and edit it once it’s auto-generated).

Anyway, I have to say it was quite fascinating to watch it automatically build multiple packages for different distros after each spec file edit, and make them automatically available in the repository once the build is finished. This site is way, way cool!

Separate your stuff!

Kohei Yoshida2 Comments

So, here is another new thing in ooo-build trunk that I just checked in. This may be particularly of interest to users in English-speaking countries.

This new option page named Formula lets you configure separators in your formula expressions. This comes in handy when, for instance, you want to separate your function parameters by commas (,) instead of semicolons (;) so that, instead of typing

=SUM(A1;B1;C1)

you could type

=SUM(A1,B1,C1)

like you are used to if you are coming from the other major spreadsheet application from the certain big company on the west coast of the US.

Likewise, you can also change the column and row separators for in-line arrays. Previously, an in-line array used semicolons (;) as the column separators and the pipe symbols (|) as the row separators, so a typical in-line array expression looked like this

={1;2;3;4;5|6;7;8;9;10}

for a 5 x 2 matrix array. By changing the column separators to commas (,) and the row separators to semicolons (;), the same expression will look like this

={1,2,3,4,5;6,7,8,9,10}

Now, the formula syntax option was there before I checked this piece in, but it was in the Calculate option page. I moved it to the new Formula page because I felt that those two configuration options really should belong together.

Anyway, this is probably good news especially for users in the English locales, where people are most used to having commas as the function argument separators but are frustrated by Calc because Calc forced them to get used to using semicolons to separate the parameters. Well, frustrate no more. :-)

Of course, those who prefer Calc’s conventional separators can revert back to them since they are configurable. So, these is no loss for the old-time Calc users who just want the old separators back. ;-)

Hiding data from chart

Kohei Yoshida2 Comments

Here is something new in ooo-build trunk. You can now hide arbitrary data points in chart if their source cells are hidden in a spreadsheet document. Here is the proof:

There is now a new check box in the Data Series Option page, to select whether or not to plot data from hidden cells. By default, chart plots all data points regardless of cell’s visibility (which means the check box is on by default). By un-selecting the check box, the chart now only plots data from visible cells, and every time you show or hide a part of the source data range, the chart gets updated.

The upstreaming effort is underway in the koheichart01 CWS. I’m trying to squeeze this into 3.1, but because of the limited availability of QA resource, it might likely slip into 3.2. I’m still trying to find someone who can QA my CWS, but let’s see what happens.