[R] Finding nrows with specefic values&converting a matrix into a table

David L Carlson dcarlson at tamu.edu
Sun Apr 30 18:33:34 CEST 2017


Show us the code you used. Don't just tell us what you did. It is likely that something you did after creating the matrix converted it to a data frame. Copy and paste your code to your emails.

> str(mydf)
'data.frame':   3 obs. of  3 variables:
 $ x: int  0 NA NA
 $ y: int  5 0 NA
 $ z: int  67 23 0

> data.frame(as.table(as.matrix(mydf)))
  Var1 Var2 Freq
1    x    x    0
2    y    x   NA
3    z    x   NA
4    x    y    5
5    y    y    0
6    z    y   NA

David C


From: abo dalash [mailto:abo_dlsh at hotmail.com] 
Sent: Sunday, April 30, 2017 11:13 AM
To: David L Carlson <dcarlson at tamu.edu>
Subject: Re: [R] Finding nrows with specefic values&converting a matrix into a table

str(mymatrix)
The Structure shows that this is a 'data.frame'?  of 120 obs. and 120 variables 
of numeric type. R deals with my matrix as a data frame although I used 
the function matrix() to produce this matrix which is not clear to me why. As this is already a data.frame, this may explains why R returns me the same 
matrix. What do you recommend now?

Many thanks 


________________________________________
From: David L Carlson <dcarlson at tamu.edu>
Sent: 30 April 2017 06:47 PM
To: abo dalash; r-help at R-project.org
Subject: RE: [R] Finding nrows with specefic values&converting a matrix into a table 
 
You did not give me any information about about your data using str() or class() so I'll guess that you have a matrix, e.g.:

> class(moredata)
[1] "matrix"
> as.data.frame.table(moredata)
  Var1 Var2 Freq
1    x    x    0
2    y    x   NA
3    z    x   NA
4    x    y    5
5    y    y    0
6    z    y   NA
7    x    z   67
8    y    z   23
9    z    z    0


David C

From: abo dalash [mailto:abo_dlsh at hotmail.com] 
Sent: Sunday, April 30, 2017 10:09 AM
To: David L Carlson <dcarlson at tamu.edu>; r-help at R-project.org
Subject: Re: [R] Finding nrows with specefic values&converting a matrix into a table

Dear David ..,

Many thanks for this detailed answer.

Your guidance reg. the first task has resolved my issue and I have understood now how to perform this type of analysis. I have saved your learning tips in my script.

Reg. the Matrix-table conversion, could you please clarify this more?.
I applied the function as.data.frame but this returned the same matrix 
without converting it into a list table. I'm not sure where is the problem
in my code :   mymatrix <- as.data.frame(mymatrix).

Many thanks for your support

Regards


________________________________________
From: David L Carlson <dcarlson at tamu.edu>
Sent: 29 April 2017 11:38 PM
To: abo dalash; r-help at R-project.org
Subject: RE: [R] Finding nrows with specefic values&converting a matrix into a table 
 
First. Do not use html messages, only plain text. Second. Provide a small example data set, preferably using dput(). Just printing your data can hide important information. Third. Read the documentation. Your first example does not return a logical vector at all:

> dput(mydata)
structure(list(Col1 = c(123L, 443L, 566L), Col2 = c(566L, 54L, 
44L), Col3 = c(235L, 566L, 235L)), .Names = c("Col1", "Col2", 
"Col3"), class = "data.frame", row.names = c(NA, -3L))

> which(mydata == 566,235)
     row col
[1,]   3   1
[2,]   1   2
[3,]   2   3

It locates cells with 566, but not 235 which is not a surprise because you did not provide a valid logical expression to which(). 

There are a number of ways to get what you want, but since you want to process rows, apply() is straightforward:

> Val566 <- apply(mydata, 1, function(x) any(x == 566))
> Val566
[1] TRUE TRUE TRUE
> Val235 <- apply(mydata, 1, function(x) any(x == 235)) 
> Val235
[1]  TRUE FALSE  TRUE
> which(Val235 & Val566)
[1] 1 3

You should read the manual pages on any(), apply(), dput() and which() and logical expressions:

> ?apply
> ?any
> ?dput
> ?which
> ?Comparison # ?"==" will also get you there.

For the second question, assuming you are beginning with a table object as R defines that term and not a matrix (since all tables are matrices, but all matrices are not tables):

> dput(moredata)
structure(c(0L, NA, NA, 5L, 0L, NA, 67L, 23L, 0L), .Dim = c(3L, 
3L), .Dimnames = list(c("x", "y", "z"), c("x", "y", "z")), class = "table")
> moredata
   x  y  z
x  0  5 67
y NA  0 23
z NA NA  0

Note, that your example uses na rather than NA. R is case sensitive so na is just an ordinary character string while NA is a missing value indicator. This is one of the reasons that dput() is important

> moredata.df <- as.data.frame(moredata)
> moredata.df
  Var1 Var2 Freq
1    x    x    0
2    y    x   NA
3    z    x   NA
4    x    y    5
5    y    y    0
6    z    y   NA
7    x    z   67
8    y    z   23
9    z    z    0
> moredata.df[order(moredata.df$Freq, decreasing=TRUE), ]
  Var1 Var2 Freq
7    x    z   67
8    y    z   23
4    x    y    5
1    x    x    0
5    y    y    0
9    z    z    0
2    y    x   NA
3    z    x   NA
6    z    y   NA

For this you should read the following manual pages:

> ?as.data.frame
> ?order
> ?Extract

-------------------------------------
David L Carlson
Department of Anthropology
Texas A&M University
College Station, TX 77840-4352

-----Original Message-----
From: R-help [mailto:r-help-bounces at r-project.org] On Behalf Of abo dalash
Sent: Saturday, April 29, 2017 10:14 AM
To: r-help at R-project.org
Subject: [R] Finding nrows with specefic values&converting a matrix into a table

Hi All


I'm trying to identify number of rows containing 2 specific values.

I tried : which(mydata == 566,235), but this returns logical values for all rows and any T in a certain row indicates the existence of one of these values but what I need to know is only number of rows in my data set with these 2 particular values considering these two values

as one pair per column. For example :


1          123   566    235

2          443    54      566

3          566    44      235


here number of rows with the values 566&235 is 2 which are

rows 1 & 3. Row 2 has only 566 so it should not be included in

our calculation.


I also have a large matrix and wanted to convert it into a table so I can

easily identify the combination with higher frequencies.


The matrix looks like this:


                    x      y      z

x                  0      5       67

y                  na    0      23

z                   na   na      0


and I would like to convert this into a table arranged with

higher values first like this :

x       z       67

y       z       23

x       y        5

x       x        0

y       y        0

z        z        0

y        x        na

z        x        na

z        y        na


Is there simple function to perform this conversion with some explanation about the Syntax


Regards



        [[alternative HTML version deleted]]

______________________________________________
R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
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