checkboxes in search options in gfilter()

[landroni]

Now the search options pops up a selection or "radio" choices. Would it be possible to have checkboxes instead? As it is in the case of grepl()?

[jverzani]

That should be easy. What choices do you want?

On Wed, Apr 10, 2013 at 4:37 PM, landroni notifications@github.com wrote:

Now the search options pops up a selection or "radio" choices. Would it be possible to have checkboxes instead? As it is in the case of grepl()?

— Reply to this email directly or view it on GitHubhttps://github.com/jverzani/gWidgets2/issues/15 .

John Verzani Chair, Department of Mathematics College of Staten Island, CUNY verzani@math.csi.cuny.edu

[jverzani]

I added ignore.case, perl and fixed.

On Wed, Apr 10, 2013 at 5:00 PM, John Verzani verzani@math.csi.cuny.eduwrote:

That should be easy. What choices do you want?

On Wed, Apr 10, 2013 at 4:37 PM, landroni notifications@github.comwrote:

Now the search options pops up a selection or "radio" choices. Would it be possible to have checkboxes instead? As it is in the case of grepl()?

— Reply to this email directly or view it on GitHubhttps://github.com/jverzani/gWidgets2/issues/15 .

John Verzani Chair, Department of Mathematics College of Staten Island, CUNY verzani@math.csi.cuny.edu

John Verzani Chair, Department of Mathematics College of Staten Island, CUNY verzani@math.csi.cuny.edu

[landroni]

This looks nice, thanks. I would rename fixed to regex and invert the resulting svalue(). I would also propose that we have regex=F and ignore.case=T by default. What do you think?

Also, is it straightforward to have a 'Whole word' option? I guess not, at least not within grepl().

[jverzani]

Okay on all but whole word search unless you have a grepl replacement to drop in. --J

On Thu, Apr 11, 2013 at 4:13 PM, landroni notifications@github.com wrote:

This looks nice, thanks. I would rename fixed to regex and invert the resulting svalue(). I would also propose that we have regex=F and ignore.case=T by default. What do you think?

Also, is it straightforward to have a 'Whole word' option? I guess not, at least not within grepl().

— Reply to this email directly or view it on GitHubhttps://github.com/jverzani/gWidgets2/issues/15#issuecomment-16258379 .

John Verzani Chair, Department of Mathematics College of Staten Island, CUNY verzani@math.csi.cuny.edu

[landroni]

Two points:

• It seems that the 'ignore case' is currently broken. Whether (de/)activated, the search is always with ignore.case=F.
• A wild idea: would it be nicer to use Hadley's stringr package for this? He provides wrappers for grepl(): str_detect(), fixed(), ignore.case() and perl(). I sent him a request for a whole.word(), but I'm still wondering if it makes sense.

[landroni]

Quick update. For some reason I keep getting this in warnings() when loading Cars93 in dffilter(): 50: In function (pattern, x, ignore.case = FALSE, perl = FALSE, ... : argument 'ignore.case = TRUE' will be ignored

Any ideas why?

[jverzani]

I'm guessing it is related to this:

fixed: logical. If ‘TRUE’, ‘pattern’ is a string to be matched as is. Overrides all conflicting arguments.

The Regex value is setting fixed=TRUE

Unfortunately, I can't disable the checkboxes without digging into the gmenu code.

On Sat, Apr 20, 2013 at 11:08 AM, landroni notifications@github.com wrote:

Quick update. For some reason I keep getting this in warnings() when loading Cars93 in dffilter(): 50: In function (pattern, x, ignore.case = FALSE, perl = FALSE, ... : argument 'ignore.case = TRUE' will be ignored

Any ideas why?

— Reply to this email directly or view it on GitHubhttps://github.com/jverzani/gWidgets2/issues/15#issuecomment-16708544 .

John Verzani Chair, Department of Mathematics College of Staten Island, CUNY verzani@math.csi.cuny.edu

[landroni]

Indeed, but that's weired. I don't understand the reasoning for this:

grepl('A', letters, fixed=T, ignore.case=T) [1] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE [17] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE Warning message: In grepl("A", letters, fixed = T, ignore.case = T) : argument 'ignore.case = TRUE' will be ignored

And strangely stringr takes a similar approach:

str_detect(letters, ignore.case(fixed('A'))) [1] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE [17] FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE Warning message: In grepl("A", c("a", "b", "c", "d", "e", "f", "g", "h", "i", "j", : argument 'ignore.case = TRUE' will be ignored

[landroni]

Since 'ignore.case = TRUE' and fixed=TRUE are incompatible, the current defaults feel buggy: "Ignore case" is checked, but filtering is case sensitive. Before a more stable solution is found, I would propose to have 'Regex' checked too by default. What do you think?

[jverzani]

Okay, changed this and pushed to github. I can't test right now. Since upgrading to 3.0.0 my Gtk setup is very flaky. Am waiting for a fix.

On Thu, May 23, 2013 at 4:03 PM, landroni notifications@github.com wrote:

Since 'ignore.case = TRUE' and fixed=TRUE are incompatible, the current defaults feel buggy: "Ignore case" is checked, but filtering is case sensitive. Before a more stable solution is found, I would propose to have 'Regex' checked too by default. What do you think?

— Reply to this email directly or view it on GitHubhttps://github.com/jverzani/gWidgets2/issues/15#issuecomment-18368166 .

John Verzani Chair, Department of Mathematics College of Staten Island, CUNY verzani@math.csi.cuny.edu

[landroni]

"Okay, changed this and pushed to github." Thanks. I tested but there seems to be a bug somewhere. If you launch dffilter(), add item, Manufacturer, then type 'ac' and notice that 'Acura' is not displayed. Then uncheck and re-check 'Regex' and 'Acura' will also be displayed. I looked into the gfilter() code and couldn't find any obvious issue there. Maybe something wrong with initialization?

"Since upgrading to 3.0.0 my Gtk setup is very flaky." What do you mean? I have both Gtk3 and Gtk2 here, without obvious issues sofar.

[jverzani]

Sorry, I meant since upgrading R, there is a serious bug with RGtk2 that keeps me from doing anything.

Will have to check some other way if this is still an issue. All I changed was the default value for the checkbox.

On Sat, May 25, 2013 at 3:41 PM, landroni notifications@github.com wrote:

"Okay, changed this and pushed to github." Thanks. I tested but there seems to be a bug somewhere. If you launch dffilter(), add item, Manufacturer, then type 'ac' and notice that 'Acura' is not displayed. Then uncheck and re-check 'Regex' and 'Acura' will also be displayed. I looked into the gfilter() code and couldn't find any obvious issue there. Maybe something wrong with initialization?

"Since upgrading to 3.0.0 my Gtk setup is very flaky." What do you mean? I have both Gtk3 and Gtk2 here, without obvious issues sofar.

— Reply to this email directly or view it on GitHubhttps://github.com/jverzani/gWidgets2/issues/15#issuecomment-18452732 .

John Verzani Chair, Department of Mathematics College of Staten Island, CUNY verzani@math.csi.cuny.edu

[jverzani]

Okay, think I fixed it. Sorry about that.

On Sat, May 25, 2013 at 3:41 PM, landroni notifications@github.com wrote:

"Okay, changed this and pushed to github." Thanks. I tested but there seems to be a bug somewhere. If you launch dffilter(), add item, Manufacturer, then type 'ac' and notice that 'Acura' is not displayed. Then uncheck and re-check 'Regex' and 'Acura' will also be displayed. I looked into the gfilter() code and couldn't find any obvious issue there. Maybe something wrong with initialization?

"Since upgrading to 3.0.0 my Gtk setup is very flaky." What do you mean? I have both Gtk3 and Gtk2 here, without obvious issues sofar.

— Reply to this email directly or view it on GitHubhttps://github.com/jverzani/gWidgets2/issues/15#issuecomment-18452732 .

John Verzani Chair, Department of Mathematics College of Staten Island, CUNY verzani@math.csi.cuny.edu

[landroni]

Excellent, thanks. It works now.

PS I was contemplating upgrading to R 3.0, but now you've given me a good reason not to. I'm curious, is there any RGtk3 on the horizon?

[jverzani]

The issue seems to be limited to the Mac. Likely it comes from the gtk-libraries shipped with att. When I get the time, I'll see if I can't install all the stuff from source to check if that works I know with the Gtk libraries installed with Julia that there aren't these issues.

As for RGtk3, interesting question. The bindings are generated automatically, so it may not be that much work. That being said, I'm not sure who is going to want to take that task on.

--J

On Sun, May 26, 2013 at 1:39 AM, landroni notifications@github.com wrote:

Excellent, thanks. It works now.

PS I was contemplating upgrading to R 3.0, but now you've given me a good reason not to. I'm curious, is there any RGtk3 on the horizon?

— Reply to this email directly or view it on GitHubhttps://github.com/jverzani/gWidgets2/issues/15#issuecomment-18458369 .

John Verzani Chair, Department of Mathematics College of Staten Island, CUNY verzani@math.csi.cuny.edu