Variable Labels from Hmisc and lm function R3.5.0

[bangecon]

When I updated to R 3.5.0, several lm and glm command lines started giving an error, "Error: x must be a numeric or a character vector." Running the same lines in R 3.4.4 did not give that command. I was able to determine that these commands ran fine in R 3.5.0 if I ran them before assigning variable labels using the label() command, but not after assigning the labels. My short run solution is "don't use the labels." But descriptive English labels are nice to have when you have a lot of variables.

[couthcommander]

Can you provide an example? The following works for me.

library(Hmisc)
ctl <- c(4.17,5.58,5.18,6.11,4.50,4.61,5.17,4.53,5.33,5.14)
trt <- c(4.81,4.17,4.41,3.59,5.87,3.83,6.03,4.89,4.32,4.69)
group <- gl(2, 10, 20, labels = c("Ctl","Trt"))
label(group) <- "Treatment Group"
weight <- c(ctl, trt)
summary(lm(weight ~ group))

class(group) returns [1] "labelled" "factor"

[bangecon]

Well, at first I ran your example and it gave me the same error. I figured something must be fishy, so I re-installed Hmisc and re-ran your example. It worked, so I re-ran my example. It also worked. So, there must have been something that didn't update properly when I updated R and my packages. At any rate, fixed now? We'll hope? Thanks for your time attention. Jim

On Wed, May 30, 2018 at 5:48 PM Cole Beck notifications@github.com wrote:

Can you provide an example? The following works for me.

library(Hmisc) ctl <- c(4.17,5.58,5.18,6.11,4.50,4.61,5.17,4.53,5.33,5.14) trt <- c(4.81,4.17,4.41,3.59,5.87,3.83,6.03,4.89,4.32,4.69) group <- gl(2, 10, 20, labels = c("Ctl","Trt")) label(group) <- "Treatment Group" weight <- c(ctl, trt) summary(lm(weight ~ group))

class(group) returns [1] "labelled" "factor"

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/harrelfe/Hmisc/issues/93#issuecomment-393344302, or mute the thread https://github.com/notifications/unsubscribe-auth/AQOL2a6-j2SdnUj431-VG_PnMKLAp82zks5t3yHCgaJpZM4UUEBM .

-- James T. Bang, Ph.D. Associate Professor and Chair Department of Economics St. Ambrose University 540-460-5948 (c)

"I was into data before it was big."

[couthcommander]

Reinstalling and restarting R is always a great place to start. My first thought was that you might have an additional package loaded interfering with behaviour. For instance, the haven package creates a labelled class that would likely cause problems if loaded after Hmisc.

[spgarbet]

Here's an example with an Hmisc dataset that was labelled showing the maven problem.

> load("data/pbc.rda")
> head(pbc)
  bili albumin stage protime    sex fu.days      age spiders hepatom ascites alk.phos   sgot chol
1 14.5    2.60     4    12.2 female     400 58.76523 present present present   1718.0 137.95  261
2  1.1    4.14     3    10.6 female    4500 56.44627 present present  absent   7394.8 113.52  302
3  1.4    3.48     4    12.0   male    1012 70.07255  absent  absent  absent    516.0  96.10  176
4  1.8    2.54     4    10.3 female    1925 54.74059 present present  absent   6121.8  60.63  244
5  3.4    3.53     3    10.9 female    1504 38.10541 present present  absent    671.0 113.15  279
6  0.8    3.98     3    11.0 female    2503 66.25873  absent present  absent    944.0  93.00  248
  trig platelet            drug status                          edema copper
1  172      190         placebo      1 edema despite diuretic therapy    156
2   88      221         placebo      0                       no edema     54
3   55      151         placebo      1     edema, no diuretic therapy    210
4   92      183         placebo      1     edema, no diuretic therapy     64
5   72      136 D-penicillamine      0                       no edema    143
6   63       NA D-penicillamine      1                       no edema     50
> library(haven)
> head(pbc)
Error: `x` and `labels` must be same type
> sessionInfo()
R version 3.4.4 (2018-03-15)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04 LTS

Matrix products: default
BLAS: /usr/lib/x86_64-linux-gnu/atlas/libblas.so.3.10.3
LAPACK: /usr/lib/x86_64-linux-gnu/atlas/liblapack.so.3.10.3

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               LC_TIME=en_US.UTF-8       
 [4] LC_COLLATE=en_US.UTF-8     LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8   
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                  LC_ADDRESS=C              
[10] LC_TELEPHONE=C             LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] haven_1.1.2

loaded via a namespace (and not attached):
 [1] compiler_3.4.4  magrittr_1.5    hms_0.4.2       tools_3.4.4     pillar_1.3.0   
 [6] rstudioapi_0.7  tibble_1.4.2    crayon_1.3.4    Rcpp_0.12.18    forcats_0.3.0  
[11] pkgconfig_2.0.2 rlang_0.2.2