How to subset() from data frame using specific rows

testdata <- c("A1", "A2", "A3", "B1", "B2", "B3")
grep("^A", testdata)

grepl("^A", testdata)

?I have a data frame called chemdata with this structure:

str(chemdata)

'data.frame': ? 14886 obs. of ?4 variables:
?$ site ? ?: Factor w/ 148 levels "BC-0.5","BC-1",..: 104 145 126 115 114
128 124 2 3 3 ...
?$ sampdate: Date, format: "1996-12-27" "1996-08-22" ...
?$ param ? : Factor w/ 8 levels "As","Ca","Cl",..: 1 1 1 1 1 1 1 1 1 1 ...
?$ quant ? : num ?0.06 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 ...

?I've looked in the R Cookbook and Dalgaard's intro book without finding a
way to use wildcards (e.g., like "BC-*") or explicitly witing each site ID
when subdsetting a data frame..

?I need to create subsets (as data frames) based on sites, but including
all sites on each stream. For example, using the initial site factor shown
above, I want a subset containing all data for sites "BC-0.5", "BC-1".
"BC-2", "BC-3", "BC-4", "BC-5", and "BC-6".

Pointers appreciated,

Rich

?I have a data frame called chemdata with this structure:

str(chemdata)

'data.frame': ? 14886 obs. of ?4 variables:
?$ site ? ?: Factor w/ 148 levels "BC-0.5","BC-1",..: 104 145 126 115 114
128 124 2 3 3 ...
?$ sampdate: Date, format: "1996-12-27" "1996-08-22" ...
?$ param ? : Factor w/ 8 levels "As","Ca","Cl",..: 1 1 1 1 1 1 1 1 1 1 ...
?$ quant ? : num ?0.06 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 ...

?I've looked in the R Cookbook and Dalgaard's intro book without finding a
way to use wildcards (e.g., like "BC-*") or explicitly witing each site ID
when subdsetting a data frame..

?I need to create subsets (as data frames) based on sites, but including
all sites on each stream. For example, using the initial site factor shown
above, I want a subset containing all data for sites "BC-0.5", "BC-1".
"BC-2", "BC-3", "BC-4", "BC-5", and "BC-6".

Pointers appreciated,

Rich

You can use something like this:

testdata <- c("A1", "A2", "A3", "B1", "B2", "B3")
grep("^A", testdata)

[1] 1 2 3

grepl("^A", testdata)

[1]  TRUE  TRUE  TRUE FALSE FALSE FALSE

On Tue, 4 Oct 2011, Sarah Goslee wrote:

You can use something like this:

testdata <- c("A1", "A2", "A3", "B1", "B2", "B3")
grep("^A", testdata)

[1] 1 2 3

grepl("^A", testdata)

[1] ?TRUE ?TRUE ?TRUE FALSE FALSE FALSE

Sarah,

?I don't see how this gives me a data frame containing only those sites I
specify. I want to plot by sites-within-streams specifying which param
factor to use.

This isn't going to be the most elegant, but it should work:
## Get the factors as characters
ff <- as.character(chemdata$site)

## Identify those that match what you want

ff <- grepl(ff, "BC-")

On Tue, 4 Oct 2011, R. Michael Weylandt wrote:

This isn't going to be the most elegant, but it should work:
## Get the factors as characters
ff <- as.character(chemdata$site)

## Identify those that match what you want

ff <- grepl(ff, "BC-")

Michael,

?Apparently grep works differently in R than it does on the command line:

bf <- grep(ff, "BC-")
Warning message:
In grep(ff, "BC-") :
?argument 'pattern' has length > 1 and only the first element will be used

?I understand what you suggest but it does not appear to work for me.

Thanks,

Rich

You asked for pointers, and didn't provide a reproducible example, so I
offered a pointer.

No, that was just a typo on my end:
the correct order of arguments should have been
ff <- grepl("BC-", ff)

On Tue, 4 Oct 2011, Sarah Goslee wrote:

You asked for pointers, and didn't provide a reproducible example, so 

offered a pointer.

Sarah,

   I did not realize that your pointer was to the factor component of 

subset() command.

   I think the most parsimonious thing for me to do is to modify the 

table with a new column of the full stream name, then re-export and 

into R.

From what we know it is difficult to say if there is some common feature 

Thanks,

Rich

Hm. I seldom use such approach. In your original request you said you want
split your data to smaller data frames based on sites

From what we know it is difficult to say if there is some common feature
in site variable. If it is organised like
XY-N
you can simply make new variable from first two letters

sites <- substr(chemdata$site,1,2)

then you can split your data frame according to sites

chem.spl <- split(chemdata, sites)

and do anything with your splitted data frames organised in list

 First thing this morning I'm upgrading to 2.13.2 and hoping that this
fixes an issue that just showed up yesterday afternoon: not being able to
access function help pages. For example, I tried ?subset and ?split because
I thought the latter is really what I want, yet R told me no help was found.
Strange; it was there a week ago.

On Wed, 5 Oct 2011, Petr PIKAL wrote:

Hm. I seldom use such approach. In your original request you said you 

split your data to smaller data frames based on sites

Petr,

   I need the additional information in the database, too.

iris.spl<- split(iris, iris$Species)
str(iris.spl)

From what we know it is difficult to say if there is some common 

in site variable. If it is organised like
XY-N
you can simply make new variable from first two letters

   Unfortunately, the site designations are not so uniform. As I went 

the process of re-doing the data I discovered this lack of consistency
resulting in duplicate records because one site had been designated XX-n 

XXn. Had to clean those up, too.

sites <- substr(chemdata$site,1,2)

then you can split your data frame according to sites

chem.spl <- split(chemdata, sites)

and do anything with your splitted data frames organised in list

   First thing this morning I'm upgrading to 2.13.2 and hoping that this
fixes an issue that just showed up yesterday afternoon: not being able 

access function help pages. For example, I tried ?subset and ?split 

I thought the latter is really what I want, yet R told me no help was 

Strange; it was there a week ago.

Thanks,

Rich

But you do not loose them, your data frame is cut according to sites
variable and put into a list