[R] stupid thing with the correlation

[Duncan Murdoch]

On 30/01/2015 5:22 AM, n omranian via R-help wrote:
> Hi All,
> I'm getting actually nuts. I don't understand the following lines in R:
> Here is the data, all rows are exactly the same!
>> ord_data[pid,]
>              c0m2     c0m4     c0m8  c0m16    c0m24    c0m48     c0p2     c0p4     c0p8   c0p16    c0p24    c0p48   c24m2    c24m4    c24m8  c24m16
> 13336382 5.632195 6.442133 5.818143 5.7683 5.862075 6.533181 5.807341 6.709774 5.664199 5.54022 5.385181 6.531977 5.29061 5.776121 6.136176 6.34699
> 13465320 5.632195 6.442133 5.818143 5.7683 5.862075 6.533181 5.807341 6.709774 5.664199 5.54022 5.385181 6.531977 5.29061 5.776121 6.136176 6.34699
> 13467455 5.632195 6.442133 5.818143 5.7683 5.862075 6.533181 5.807341 6.709774 5.664199 5.54022 5.385181 6.531977 5.29061 5.776121 6.136176 6.34699
> 13518680 5.632195 6.442133 5.818143 5.7683 5.862075 6.533181 5.807341 6.709774 5.664199 5.54022 5.385181 6.531977 5.29061 5.776121 6.136176 6.34699
>            c24m24   c24m48    c24p2    c24p4    c24p8   c24p16  c24p24   c24p48
> 13336382 5.489161 6.043948 5.807802 5.761756 5.833559 5.293438 5.92068 5.759683
> 13465320 5.489161 6.043948 5.807802 5.761756 5.833559 5.293438 5.92068 5.759683
> 13467455 5.489161 6.043948 5.807802 5.761756 5.833559 5.293438 5.92068 5.759683
> 13518680 5.489161 6.043948 5.807802 5.761756 5.833559 5.293438 5.92068 5.759683
> 
> Now I found correlation:
>> pcor <- cor(t((ord_data[pid,])))
>> pcor
>          13336382 13465320 13467455 13518680
> 13336382        1        1        1        1
> 13465320        1        1        1        1
> 13467455        1        1        1        1
> 13518680        1        1        1        1
> But, then I get this funny result !!!
> all(pcor[,1]==1)
> [1] FALSE
>> pcor[2,2]==1
> [1] TRUE
>> pcor[3,2]==1
> [1] FALSE
> Could anybody please comment on this?Many thanks.
> 	[[alternative HTML version deleted]]

Please just send your messages to r-help, not all those other addresses,
and please don't reply to an existing thread.

Your answer is given in FAQ 7.31.

Duncan Murdoch