R help - Jun 2007 - How to add elements to a table

Hi,

I've been using R for a few weeks now, but consider myself still a  
newbie, especially in what concerns the basics of data handling in R.

My situation is this:
I read in data from a database using RODBC, and then use "table" to  
produce a table of "counts per years", table which I give to
"plot".
All is well, as long as there are no gaps in the years sequence, but  
now I have the following result (example):

1990 20
1991 15
1995 17
1997  9
1998 10
1999 11
2000  5

The "plot" function is quite intelligent, in the sense that it draws  
appropriate gaps in the x-axis between years, but not intelligent  
enough to interrupt the data line during the gap. This gives the  
impression that, although no year is marked on the x-axis between 1991  
and 1995, there has been data for this period, which is not correct.

What I tried to do is convert the table to a matrix, insert zeros for  
the missing years using rbind and cbind, and convert the result back  
to table. But the structure of this resulting table is not the same as  
for the originating table, so that I need to pass "tab[1,]" to
"plot".
It's no longer a contingency table in fact.

I've seen in the mailing list archives that there is an issue on using  
"table"s when matrixes or other structures would be more appropriate.

I like the "table", because "plot" automatically plots the  
corresponding years on the x-axis, which I find less error-prone than  
adding the tick labels later by hand, i.e. the association between  
years and counts is stronger.

Also, as I tabulate counts of cases per gender, or per age categories,  
I think a contingency table is the right thing to use, isn't it?

I'd be glad on any advice about what would be the best data structure  
to handle my situation, and I'd also like to know how I could  
automagically check a list of "year, cases" rows against a fixed list
of years, insert zero or "NA" values for missing years, and get an  
easily usable result that I can forward to "plot" or
"barplot".

Thanks a lot in advance,

Anne-Marie

> z <- read.table("clipboard")
 > z
    V1 V2
1 1990 20
2 1991 15
3 1995 17
4 1997  9
5 1998 10
6 1999 11
7 2000  5
 > zz <- merge(data.frame(V1=1990:2000), z, by="V1", all.x=T)
 > plot(zz, type="l")

Jacques VESLOT

INRA - Biostatistique & Processus Spatiaux
Site Agroparc 84914 Avignon Cedex 9, France

Tel: +33 (0) 4 32 72 21 58
Fax: +33 (0) 4 32 72 21 84



bioinfonews at pt.lu a ?crit :
> Hi,
>
> I've been using R for a few weeks now, but consider myself still a  
> newbie, especially in what concerns the basics of data handling in R.
>
> My situation is this:
> I read in data from a database using RODBC, and then use "table"
to
> produce a table of "counts per years", table which I give to
"plot".
> All is well, as long as there are no gaps in the years sequence, but  
> now I have the following result (example):
>
> 1990 20
> 1991 15
> 1995 17
> 1997  9
> 1998 10
> 1999 11
> 2000  5
>
> The "plot" function is quite intelligent, in the sense that it
draws
> appropriate gaps in the x-axis between years, but not intelligent  
> enough to interrupt the data line during the gap. This gives the  
> impression that, although no year is marked on the x-axis between 1991  
> and 1995, there has been data for this period, which is not correct.
>
> What I tried to do is convert the table to a matrix, insert zeros for  
> the missing years using rbind and cbind, and convert the result back  
> to table. But the structure of this resulting table is not the same as  
> for the originating table, so that I need to pass "tab[1,]" to
"plot".
> It's no longer a contingency table in fact.
>
> I've seen in the mailing list archives that there is an issue on using
> "table"s when matrixes or other structures would be more
appropriate.
>
> I like the "table", because "plot" automatically plots
the
> corresponding years on the x-axis, which I find less error-prone than  
> adding the tick labels later by hand, i.e. the association between  
> years and counts is stronger.
>
> Also, as I tabulate counts of cases per gender, or per age categories,  
> I think a contingency table is the right thing to use, isn't it?
>
> I'd be glad on any advice about what would be the best data structure  
> to handle my situation, and I'd also like to know how I could  
> automagically check a list of "year, cases" rows against a fixed
list
> of years, insert zero or "NA" values for missing years, and get
an
> easily usable result that I can forward to "plot" or
"barplot".
>
> Thanks a lot in advance,
>
> Anne-Marie
>
> ______________________________________________
> R-help at stat.math.ethz.ch 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.
>