Missing tick mark labels from barChart / quantmod - plot.xts missing

source("axTicksByTimeModified.r")
assignInNamespace("axTicksByTime",axTicksByTime, ns="xts") 

Dear Jeff,
I am glad to tell you that I have traced the problem into a line of code 
concerning the parameter of "%n" in time format string.
Here is the data for test:

x<-xts(x[,-1],as.Date(x[,1]))     #convert data frame to xts.   
head(x)

           Symbol     Close   
1990-12-19 SH000001     99.98
1990-12-20 SH000001    104.39 
1990-12-21 SH000001    109.13 
1990-12-24 SH000001    114.55 
1990-12-25 SH000001    120.25

plot(x[,"Close"],auto.grid=TRUE,minor.ticks=FALSE)

The plot is OK except that no date labels for x axis.  Major ticks and
grids are all Ok for x and y axis.

Then I used the xts:::plot.xts method  you taught me to get the definition
of plot.xts.
I found that a line of code as
ep <- axTicksByTime(x, "auto")
returned ep with no names

names(ep)

[1] "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" ""

Then I trace the axTicksByTime.  There is a line of code as:
names(ep) <- format(index(x)[ep], fmt)
The  fmt is something like "%b %d%n%Y".

ep

                                                                           
   1   10  263  514  767 1010 1259 1505 1748 1994 2233 2472 2712 2949 3190 
                         
3433 3674 3915 4157 4393 

format(index(x)[ep])

 [1] "1990-12-19" "1991-01-02" "1992-01-02" "1993-01-04" "1994-01-03"
"1995-01-03" "1996-01-02" "1997-01-02" "1998-01-05"
[10] "1999-01-04" "2000-01-04" "2001-01-02" "2002-01-04" "2003-01-02"
"2004-01-02" "2005-01-04" "2006-01-04" "2007-01-04"
[19] "2008-01-02" "2008-12-18"

format(index(x)[ep], "%b %d%n%Y")

 [1] "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" ""

I test with several definitions of format string, the format() will output
nothing with "%n" in the format string.   If I remove "%n", everything is
OK.

The help said that "%n" is for a  new line.
I guess maybe "%n" just conflict with something in my computer.

Since my vista is in Chinese, the "%b" will output month names in Chinese. 
I omit the "%b".  The problem is still there.

format(index(x)[ep], "%d%n%Y")

 [1] "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" ""

Then I changed the locale to US:

Sys.setlocale("LC_TIME","US")

[1] "English_United States.1252"

format still has problems with "%n".

format(index(x)[ep], "%d%n%Y")

 [1] "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" "" ""

format(index(x)[ep], "%b %Y")

 [1] "Dec 1990" "Jan 1991" "Jan 1992" "Jan 1993" "Jan 1994" "Jan 1995"
"Jan 1996" "Jan 1997" "Jan 1998" "Jan 1999"
[11] "Jan 2000" "Jan 2001" "Jan 2002" "Jan 2003" "Jan 2004" "Jan 2005"
"Jan 2006" "Jan 2007" "Jan 2008" "Dec 2008"

I could not trace the porblem futher into the definition of format()
because of my limited knowledge of computer and R :(

I guess maybe you have told me the issue since you have said:
"not showing line breaks in the date (windows doesn't recognize this)."
Previously I did not understand your comment well because I am not
familiar with R.  

If that's the case, I wonder whether there is some easy way for me to
alter the code in axTicksByTime a little in my computer to make it worked.
Thanks. 

Information on Package xts:
URL:           http://r-forge.r-project.org/projects/xts/
Packaged:      Tue Nov 18 09:15:14 2008; jryan
Built:         R 2.7.2; i386-pc-mingw32; 2008-11-18 19:22:56; windows

Information on my computer:
Fujitsu Notebook
Windows Vista Basic, OEM, Chinese.
R 2.7.2     (Previously I have got the same problem in R 2.8)

Regards,
Wind


Jeff Ryan wrote:

Hi Wind,

plot.xts is there, just not exported.  This means it gets called when
an xts object is passed to plot()

xts:::plot.xts

That aside, the issue is somewhere else.  I have tried the CRAN
release, the R-forge package, and the dev code on R-forge and all seem
to show an x axis with dates.

There may be a formatting issue with Windows and the release you have
not showing line breaks in the date (windows doesn't recognize this).
Try using this version (though probably won't fix either):

install.packages("xts",repos="http://R-Forge.R-project.org")

chartSeries and plot.xts use axTicksByTime to format the labels.  The
issue most likely is that of the line breaks.

All of the above have been fixed in not yet released /dev branch of
xts.  We are hoping to have that ready in the next few weeks.

Jeff

On Sun, Nov 16, 2008 at 3:24 AM, ? <windspeedo at qq.com> wrote:

Dear Jeff and R experts,
I wonder whether you have received the previous email on this matter.
I have narrow the problem to plot.xts missing.

require(quantmod)

Loading required package: quantmod
Loading required package: xts
Loading required package: zoo
Attaching package: 'zoo'
       The following object(s) are masked from package:base :
        as.Date.numeric
Loading required package: Defaults
quantmod: Quantitative Financial Modelling Framework
Version 0.3-6, Revision 433
http://www.quantmod.com

getSymbols('AAPL')

[1] "AAPL"

barChart(AAPL)

#successful.  But NO date label for x axis labels.

a<-Cl(AAPL)
class(a)

[1] "xts" "zoo"

plot.zoo(a)

#successful.  There are annual label for x axis major ticks.

chartSeries(a)

Error in sum(poss.new) : invalid 'type' (list) of argument

plot(a)

#no date label for x  axis ticks

plot(as.xts(a))

#same graph as the above. still no date label for x axis ticks

plot.xts(a)

Error: could not find function "plot.xts"

?plot.xts

#the help window for plot.xts poped up.

library(help=xts)

#window poped up with the following content:
Version:       0.0-15
Revision:      139
Date:          2007-06-04
Author:        Jeffrey A. Ryan, Josh M. Ulrich
Depends:       zoo
...
License:       GPL-3
URL:           http://r-forge.r-project.org/projects/xts/
Packaged:      Mon Jun 9 23:47:46 2008; jryan
Built:         R 2.7.2; i386-pc-mingw32; 2008-10-15 23:19:56; windows
Index:
...
plot.xts                Plotting xts Objects
...

It seems that plot.zoo does its job but plot.xts may have some problem.

Besides, the OS is Vista Oem on a notebook.  R is 2.7.2 or 2.8.0.

I wonder whether it is enough for R experts to diagnose the problem.
Thanks.

Wind

-- 
Jeffrey Ryan
jeffrey.ryan at insightalgo.com

ia: insight algorithmics
www.insightalgo.com