0

I am trying to convert my data from wide to long. I have an id variable and 131 columns of time data. All columns are integer format. The data was imported in csv format.

> class(ECGHR)
[1] "tbl_df"     "tbl"        "data.frame"

Here are the first 6 rows of data for 22 time-varying variables:

> head (ECGHR)
# A tibble: 6 × 132
 id    T0    T1    T2    T3    T4    T5    T6    T7    T8    T9   T10   T11  T12   T13   T14   T15   T16   T17   T18   T19   T20   T21   T22
<int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int>
1  2003    70    65    69    71    68    74    67    65    60    66    59    99    74    73    87    94    88    88    98    73    72    63    79
2  2004    98   105    85    87    83    83    77    71    75    73    73    71    87    74    79    75    72    71    71    68    86    72    73
3  2008    93    92    91    90    93    88    89    83    93    96    95    94    89    91    87    93    88   100    93    92    89    93    83
4  2010    71    69    74    71    74    68    74    68    78    75    81    72    74    77    74    70    64    68    65    69    66    73    75
5  2018    90    86    93    91    86    82    97    95    95    83    79    88    78    92    86    94    83    79    96    91    72    78    80
6  2019    60    58    59    58    66    68    73    77    58   110    56    56    53    49    73    56    45    50    46    45    46    54    50

Here is my reshape code: 

ECGHR_long=reshape(ECGHR, varying=c(2:132), direction="long", idvar="id", sep="")

Here is the warning message:

Error in 'row.names<-.data.frame'('tmp', value = paste(d[, idvar], times[1L], : invalid 'row.names' length In addition: Warning message: Setting row names on a tibble is deprecated.

I am used to working with data frames and I can't work out why it won't reshape (I'm new to R). Do I need to convert it to a dat frame? I'd be very grateful for any insights!

eli-k
  • 10,898
  • 11
  • 40
  • 44
sara_khan
  • 21
  • 6

1 Answers1

3

To use the reshape function in the base of R try this -- we have omitted idvar="id" since it can figure that one out itself but it won't hurt if you add it in.

reshape(as.data.frame(ECGHR), dir = "long", varying = list(2:ncol(ECGHR)),
  times = names(ECGHR)[-1])
##           id time T0
## 2003.T0 2003   T0 70
## 2004.T0 2004   T0 98
## 2008.T0 2008   T0 93
## 2010.T0 2010   T0 71
## 2018.T0 2018   T0 90
## 2019.T0 2019   T0 60
## etc.

or using the reshape2 package:

library(reshape2)
melt(ECGHR, id.vars = 1)

##     id variable value
## 1 2003       T0    70
## 2 2004       T0    98
## 3 2008       T0    93
## 4 2010       T0    71
## 5 2018       T0    90
## 6 2019       T0    60

Note: We used this input:

Lines <- "id    T0    T1    T2    T3    T4    T5    T6    T7    T8    T9   T10   T11  T12   T13   T14   T15   T16   T17   T18   T19   T20   T21   T22
1  2003    70    65    69    71    68    74    67    65    60    66    59    99    74    73    87    94    88    88    98    73    72    63    79
2  2004    98   105    85    87    83    83    77    71    75    73    73    71    87    74    79    75    72    71    71    68    86    72    73
3  2008    93    92    91    90    93    88    89    83    93    96    95    94    89    91    87    93    88   100    93    92    89    93    83
4  2010    71    69    74    71    74    68    74    68    78    75    81    72    74    77    74    70    64    68    65    69    66    73    75
5  2018    90    86    93    91    86    82    97    95    95    83    79    88    78    92    86    94    83    79    96    91    72    78    80
6  2019    60    58    59    58    66    68    73    77    58   110    56    56    53    49    73    56    45    50    46    45    46    54    50"

ECGHR <- read.table(text = Lines, header = TRUE)
G. Grothendieck
  • 254,981
  • 17
  • 203
  • 341
  • Thank you for your suggestion. I get the same error message as posted in my question when I use reshape. – sara_khan Apr 05 '17 at 11:12
  • Try replacing `ECGHR` with `as.data.frame(ECGHR)` in case `reshape` does not like tibbles -- as per modified code in answer. At any rate the code in my answer works when applied to the input in the Note at its end as you can see from the output produced so try to determine what is different. – G. Grothendieck Apr 05 '17 at 11:33