R help - Aug 2011 - gsub for numeric characters in string

Dear all,

I have what is a bit of a confusing question, so I hope that I can explain
clearly. Thank you for your help in advance.

I would like to do a replacement procedure on several strings, but the way
that I am currently going about it is not working.

I have defined "len", which is a series comprising the lengths of
different
items, all preceded by a colon.
> len
[1] ":328" ":154"  ":135"
[4] ":147" ":30"  ":50"
[7] ":252" ":45"  ";"

'lenplustate" is a series that is comprised of an attribute of each
item
preceding the colon, followed by the length (as defined in len).
> lenplusstate
[1] "1:328"  "1:154"   "4:135"
[4] "NA:147" "3:30"   "2:50"
[7] "NA:252" "NA:45"  "NA;"

"tree" is a string that gives the specific relationship (via
parenthetical
notation) among the different items. Note that the lengths are included in
this tree (following the colon), and the name of each item (1-5) precedes a
colon. However, not every colon is preceded with a name (because there are
internal nodes in the tree structure).
> tree
[1] "(*1*:328,((*5*:154,*2*:135):147,(*3*:30,*4*:50):252):45);"

I would like to replace the length with the lengthplusstate in the tree,
while removing the names, so that it looks like this:
> theoreticalnewtree
[1] "(*1*:328,((*1*:154,*4*:135)NA:147,(*3*:30,*2*:50)NA:252)NA:45);"

I am using this code:

 > for (j in all)
            newtree <- gsub(ln[j], lnplusstate[j], tree)

However, I end up with this:
> newtree
[1]
"(*11*:328,((*51*:154,*24*:135)NA:147,(*33*:30,*42*:50)NA:252)NA:45);"


that is, I have not removed the names from the string so that now the state
information is wrong.

If anyone can help with the proper code to get rid of the names in tree
while replacing the length with lengthplustate, I would appreciate it.

I apologize if this was unclear.

All the best,

Rebecca

	[[alternative HTML version deleted]]

On Fri, Aug 19, 2011 at 11:11 AM, Rebecca Gray <atlasrrg at gmail.com>
wrote:
> Dear all,
>
> I have what is a bit of a confusing question, so I hope that I can explain
> clearly. Thank you for your help in advance.
>
> I would like to do a replacement procedure on several strings, but the way
> that I am currently going about it is not working.
>
> I have defined "len", which is a series comprising the lengths of
different
> items, all preceded by a colon.
>
>> len
> [1] ":328" ":154" ?":135"
> [4] ":147" ":30" ?":50"
> [7] ":252" ":45" ?";"
>
> 'lenplustate" is a series that is comprised of an attribute of
each item
> preceding the colon, followed by the length (as defined in len).
>
>> lenplusstate
> [1] "1:328" ?"1:154" ? "4:135"
> [4] "NA:147" "3:30" ? "2:50"
> [7] "NA:252" "NA:45" ?"NA;"
>
> "tree" is a string that gives the specific relationship (via
parenthetical
> notation) among the different items. Note that the lengths are included in
> this tree (following the colon), and the name of each item (1-5) precedes a
> colon. However, not every colon is preceded with a name (because there are
> internal nodes in the tree structure).
>
>> tree
> [1] "(*1*:328,((*5*:154,*2*:135):147,(*3*:30,*4*:50):252):45);"
>
> I would like to replace the length with the lengthplusstate in the tree,
> while removing the names, so that it looks like this:
>
>> theoreticalnewtree
> [1]
"(*1*:328,((*1*:154,*4*:135)NA:147,(*3*:30,*2*:50)NA:252)NA:45);"
>
I can help you, but what is the "name" of each item? I thought it was
the index of the item in the len and lenplusstate variables, but that
apparently is not the case. You have to specify the names as well.

Peter