[R] format.pval () and printCoefmat ()
Muhuri, Pradip (SAMHSA/CBHSQ)
Pradip.Muhuri at samhsa.hhs.gov
Sat Dec 15 19:19:43 CET 2012
Hi Arun,
Thank you so much for further clarifications and help.
Pradip
Pradip K. Muhuri, PhD
Statistician
Substance Abuse & Mental Health Services Administration
The Center for Behavioral Health Statistics and Quality
Division of Population Surveys
1 Choke Cherry Road, Room 2-1071
Rockville, MD 20857
Tel: 240-276-1070
Fax: 240-276-1260
e-mail: Pradip.Muhuri at samhsa.hhs.gov
The Center for Behavioral Health Statistics and Quality your feedback. Please click on the following link to complete a brief customer survey: http://cbhsqsurvey.samhsa.gov
-----Original Message-----
From: arun [mailto:smartpink111 at yahoo.com]
Sent: Saturday, December 15, 2012 11:04 AM
To: Muhuri, Pradip (SAMHSA/CBHSQ)
Cc: R help; David Winsemius
Subject: Re: [R] format.pval () and printCoefmat ()
Hi Pradip,
It this is just formatting issue, it is possible to do that with ?formatC() or ?sprintf(), but it may change those variables from numeric to character.
One possibilty from `res`:
res<-data.frame(dat1[,1:2],read.table(text=Lines2,header=TRUE))
varsNum<-sapply(res,is.numeric)
res[varsNum]<-lapply(res[varsNum],round,digits=3)
#Here, the numeric columns with digits<3 are not changed, but the ones with >3 were all changed to digits3.
As I mentioned, sprintf() changes the number of digits
as.data.frame(do.call(cbind,lapply(res[varsNum],function(x) sprintf("%.3f",x))))
# mean_level1 mean_level2 rel_diff p_mean cohens_d
#1 18.700 11.910 0.574 0.000 0.175
#2 18.700 14.460 0.297 0.000 0.110
#3 18.700 13.540 0.384 0.000 0.134
A.K.
----- Original Message -----
From: "Muhuri, Pradip (SAMHSA/CBHSQ)" <Pradip.Muhuri at samhsa.hhs.gov>
To: arun <smartpink111 at yahoo.com>
Cc: R help <r-help at r-project.org>; David Winsemius <dwinsemius at comcast.net>
Sent: Saturday, December 15, 2012 10:12 AM
Subject: RE: [R] format.pval () and printCoefmat ()
Dear Arun and David,
I am so grateful to you for all your help with the code. Thanks and regards, Pradip
Arun - All this is very helpful. In general, I can follow the code. I only have the following questions:
What changes in the code would be required to have 3 places after decimal for all numeric variables in the "res" data frame?
Thanks,
Pradip
####### below is the display of the data from Lines1, Lines2, and res
> head (data.frame(Lines1))
Lines1
1 mean_level1 mean_level2 rel_diff p_mean cohens_d
2 1 18.744 11.911 0.574 0.00 0.175
3 2 18.744 14.455 0.297 0.00 0.110
4 3 18.744 13.540 0.384 0.00 0.133
5 4 18.744 6.002 2.123 0.00 0.333
6 5 18.744 5.834 2.213 0.00 0.349
> head (data.frame(Lines2))
Lines2
1 mean_level1 mean_level2 rel_diff p_mean cohens_d
2 1 18.744 11.911 0.574 0.00 0.175
3 2 18.744 14.455 0.297 0.00 0.110
4 3 18.744 13.540 0.384 0.00 0.133
5 4 18.744 6.002 2.123 0.00 0.333
6 5 18.744 5.834 2.213 0.00 0.349
> head (res)
contrast_level1 contrast_level2 mean_level1 mean_level2 rel_diff p_mean cohens_d
1 wh 2+hi 18.7 11.91 0.574 0 0.175
2 wh 2+rc 18.7 14.46 0.297 0 0.110
3 wh aian 18.7 13.54 0.384 0 0.133
4 wh asan 18.7 6.00 2.123 0 0.333
5 wh blck 18.7 5.83 2.213 0 0.349
6 wh csam 18.7 7.93 1.363 0 0.279
________________________________________
From: arun [smartpink111 at yahoo.com]
Sent: Friday, December 14, 2012 10:12 PM
To: Muhuri, Pradip (SAMHSA/CBHSQ)
Cc: R help; David Winsemius
Subject: Re: [R] format.pval () and printCoefmat ()
Hi Pradip,
May be this helps:
dat1<-read.table(text="
contrast_level1 contrast_level2 mean_level1 mean_level2 rel_diff p_mean cohens_d
1 wh 2+hi 18.7 11.91 0.574 1.64e-05 0.1753
2 wh 2+rc 18.7 14.46 0.297 9.24e-06 0.1101
3 wh aian 18.7 13.54 0.384 9.01e-05 0.1335
4 wh asan 18.7 6.00 2.123 2.20e-119 0.3326
5 wh blck 18.7 5.83 2.213 0.00e+00 0.3490
6 wh csam 18.7 7.93 1.363 1.27e-47 0.2793
7 wh cub 18.7 10.85 0.728 6.12e-08 0.2025
8 wh dmcn 18.7 7.13 1.629 1.59e-15 0.2981
9 wh hisp 18.7 9.72 0.928 3.27e-125 0.2420
10 wh mex 18.7 9.60 0.952 8.81e-103 0.2420
11 wh nhpi 18.7 16.14 0.162 1.74e-01 0.0669
12 wh othh 18.7 NA NA NA NA
13 wh pr 18.7 10.47 0.791 3.64e-23 0.2131
14 wh spn 18.7 15.15 0.237 1.58e-02 0.0922
",sep="",header=TRUE,stringsAsFactors=FALSE)
Lines1<-capture.output(printCoefmat(dat1[,-c(1:2)],has.Pvalue=TRUE,eps.Pvalue=0.001))
Lines2<-gsub("\\s+$","",gsub("\\.$","",Lines1[1:15]))
res<-data.frame(dat1[,1:2],read.table(text=Lines2,header=TRUE))
#or
# res<-cbind(dat1[,1:2],read.table(text=Lines2,header=TRUE))
res
# contrast_level1 contrast_level2 mean_level1 mean_level2 rel_diff p_mean
#1 wh 2+hi 18.7 11.91 0.574 0.0000
#2 wh 2+rc 18.7 14.46 0.297 0.0000
#3 wh aian 18.7 13.54 0.384 0.0001
-------------------------------------------------------------
----------------------------------------------------------
# cohens_d
#1 0.1753
#2 0.1101
#3 0.1335
-------------------------------------------------
-------------------------------------------------
str(res)
#'data.frame': 14 obs. of 7 variables:
# $ contrast_level1: chr "wh" "wh" "wh" "wh" ...
# $ contrast_level2: chr "2+hi" "2+rc" "aian" "asan" ...
# $ mean_level1 : num 18.7 18.7 18.7 18.7 18.7 18.7 18.7 18.7 18.7 18.7 ...
# $ mean_level2 : num 11.91 14.46 13.54 6 5.83 ...
# $ rel_diff : num 0.574 0.297 0.384 2.123 2.213 ...
# $ p_mean : num 0e+00 0e+00 1e-04 0e+00 0e+00 0e+00 0e+00 0e+00 0e+00 0e+00 ...
# $ cohens_d : num 0.175 0.11 0.134 0.333 0.349 ...
A.K.
----- Original Message -----
From: "Muhuri, Pradip (SAMHSA/CBHSQ)" <Pradip.Muhuri at samhsa.hhs.gov>
To: 'David Winsemius' <dwinsemius at comcast.net>
Cc: R help <r-help at r-project.org>
Sent: Friday, December 14, 2012 5:18 PM
Subject: Re: [R] format.pval () and printCoefmat ()
Hi David,
Thank you so much for helping me with the code.
Your suggested code gives me the following results. Please see below. I don't understand why I am getting two blocks of prints (5 columns, and then 7 columns), with some columns repeated.
Regards,
Pradip
#############################
> cbind( y0410_1825_mf_alc[ 1:2],
+ printCoefmat(y0410_1825_mf_alc[ -(1:2) ], has.Pvalue=TRUE, eps.Pvalue=0.0001)
+ )
mean_level1 mean_level2 rel_diff p_mean cohens_d
1 18.744 11.911 0.574 0.00 0.175
2 18.744 14.455 0.297 0.00 0.110
3 18.744 13.540 0.384 0.00 0.133
4 18.744 6.002 2.123 0.00 0.333
5 18.744 5.834 2.213 0.00 0.349
6 18.744 7.933 1.363 0.00 0.279
7 18.744 10.849 0.728 0.00 0.203
8 18.744 7.130 1.629 0.00 0.298
9 18.744 9.720 0.928 0.00 0.242
10 18.744 9.600 0.952 0.00 0.242
11 18.744 16.135 0.162 0.17 0.067 .
12 18.744 NA NA NA NA
13 18.744 10.465 0.791 0.00 0.213
14 18.744 15.149 0.237 0.02 0.092 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
contrast_level1 contrast_level2 mean_level1 mean_level2 rel_diff p_mean cohens_d
1 wh 2+hi 18.7 11.91 0.574 1.64e-05 0.1753
2 wh 2+rc 18.7 14.46 0.297 9.24e-06 0.1101
3 wh aian 18.7 13.54 0.384 9.01e-05 0.1335
4 wh asan 18.7 6.00 2.123 2.20e-119 0.3326
5 wh blck 18.7 5.83 2.213 0.00e+00 0.3490
6 wh csam 18.7 7.93 1.363 1.27e-47 0.2793
7 wh cub 18.7 10.85 0.728 6.12e-08 0.2025
8 wh dmcn 18.7 7.13 1.629 1.59e-15 0.2981
9 wh hisp 18.7 9.72 0.928 3.27e-125 0.2420
10 wh mex 18.7 9.60 0.952 8.81e-103 0.2420
11 wh nhpi 18.7 16.14 0.162 1.74e-01 0.0669
12 wh othh 18.7 NA NA NA NA
13 wh pr 18.7 10.47 0.791 3.64e-23 0.2131
14 wh spn 18.7 15.15 0.237 1.58e-02 0.0922
Pradip K. Muhuri, PhD
Statistician
Substance Abuse & Mental Health Services Administration
The Center for Behavioral Health Statistics and Quality
Division of Population Surveys
1 Choke Cherry Road, Room 2-1071
Rockville, MD 20857
Tel: 240-276-1070
Fax: 240-276-1260
e-mail: Pradip.Muhuri at samhsa.hhs.gov
The Center for Behavioral Health Statistics and Quality your feedback. Please click on the following link to complete a brief customer survey: http://cbhsqsurvey.samhsa.gov
-----Original Message-----
From: David Winsemius [mailto:dwinsemius at comcast.net]
Sent: Friday, December 14, 2012 3:22 PM
To: Muhuri, Pradip (SAMHSA/CBHSQ)
Cc: R help
Subject: Re: [R] format.pval () and printCoefmat ()
On Dec 14, 2012, at 11:48 AM, Muhuri, Pradip (SAMHSA/CBHSQ) wrote:
> Hi List,
>
> My goal is to force R not to print in scientific notation in the sixth column (rel_diff - for the p-value) of my data frame (not a matrix).
>
> I have used the format.pval () and printCoefmat () functions on the data frame. The R script is appended below.
>
> This issue is that use of the format.pval () and printCoefmat () functions on the data frame gives me the desired results, but coerces the character string into NAs for the two character variables, because my object is a data frame, not a matrix. Please see the first output below: contrast_level1 contrast_level2).
>
> Is there a way I could have avoid printing the NAs in the character fields
They are probably factor columns.
> when using the format.pval () and printCoefmat () on the data frame?
>
> I would appreciate receiving your help.
>
> Thanks,
>
> Pradip
> setwd ("F:/PR1/R_PR1")
>
> load (file = "sigtests_overall_withid.rdata")
>
> #format.pval(tt$p.value, eps=0.0001)
>
> # keep only selected columns from the above data frame
> keep_cols1 <- c("contrast_level1", "contrast_level2","mean_level1",
> "mean_level2", "rel_diff",
> "p_mean", "cohens_d")
>
> #subset the data frame
> y0410_1825_mf_alc <- subset (sigtests_overall_withid,
> years=="0410" & age_group=="1825"
> & gender_group=="all" & drug=="alc"
> & contrast_level1=="wh",
> select=keep_cols1)
> #change the row.names
> row.names (y0410_1825_mf_alc)= 1:dim(y0410_1825_mf_alc)[1]
>
> #force
> format.pval(y0410_1825_mf_alc$p_mean, eps=0.0001)
Presumably that call will produce desired results since it is on only one column. (I'm not sure why you think format.pval contributed to your NA output.)
>
> #print the observations from the sub-data frame
> options (width=120,digits=3 )
> #y0410_1825_mf_alc
>
> printCoefmat(y0410_1825_mf_alc, has.Pvalue=TRUE, eps.Pvalue=0.0001)
Why not use `cbind.data.frame` rather than trying to get `printCoefmat` to do something it (apparently) wasn't designed to do?
cbind( y0410_1825_mf_alc[ 1:2],
printCoefmat(y0410_1825_mf_alc[ -(1:2) ], has.Pvalue=TRUE, eps.Pvalue=0.0001)
)
--
David.
>
> ####################### When format.pval () and printCoefmat () used
>
>
> contrast_level1 contrast_level2 mean_level1 mean_level2 rel_diff p_mean cohens_d
>
> 1 NA NA 18.744 11.911 0.574 0.00 0.175
> 2 NA NA 18.744 14.455 0.297 0.00 0.110
> 3 NA NA 18.744 13.540 0.384 0.00 0.133
> 4 NA NA 18.744 6.002 2.123 0.00 0.333
> 5 NA NA 18.744 5.834 2.213 0.00 0.349
> 6 NA NA 18.744 7.933 1.363 0.00 0.279
> 7 NA NA 18.744 10.849 0.728 0.00 0.203
> 8 NA NA 18.744 7.130 1.629 0.00 0.298
> 9 NA NA 18.744 9.720 0.928 0.00 0.242
> 10 NA NA 18.744 9.600 0.952 0.00 0.242
> 11 NA NA 18.744 16.135 0.162 0.17 0.067 .
> 12 NA NA 18.744 NA NA NA NA
> 13 NA NA 18.744 10.465 0.791 0.00 0.213
> 14 NA NA 18.744 15.149 0.237 0.02 0.092 .
> ---
> Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
> Warning messages:
> 1: In data.matrix(x) : NAs introduced by coercion
> 2: In data.matrix(x) : NAs introduced by coercion
>
> ####################### When format.pval () and printCoefmat () not used
>
> contrast_level1 contrast_level2 mean_level1 mean_level2 rel_diff p_mean cohens_d
> 1 wh 2+hi 18.7 11.91 0.574 1.64e-05 0.1753
> 2 wh 2+rc 18.7 14.46 0.297 9.24e-06 0.1101
> 3 wh aian 18.7 13.54 0.384 9.01e-05 0.1335
> 4 wh asan 18.7 6.00 2.123 2.20e-119 0.3326
> 5 wh blck 18.7 5.83 2.213 0.00e+00 0.3490
> 6 wh csam 18.7 7.93 1.363 1.27e-47 0.2793
> 7 wh cub 18.7 10.85 0.728 6.12e-08 0.2025
> 8 wh dmcn 18.7 7.13 1.629 1.59e-15 0.2981
> 9 wh hisp 18.7 9.72 0.928 3.27e-125 0.2420
> 10 wh mex 18.7 9.60 0.952 8.81e-103 0.2420
> 11 wh nhpi 18.7 16.14 0.162 1.74e-01 0.0669
> 12 wh othh 18.7 NA NA NA NA
> 13 wh pr 18.7 10.47 0.791 3.64e-23 0.2131
> 14 wh spn 18.7 15.15 0.237 1.58e-02 0.0922
>
>
>
> Pradip K. Muhuri, PhD
>
David Winsemius
Alameda, CA, USA
______________________________________________
R-help at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
More information about the R-help
mailing list