I want to merge 4 columns together, (Standing, Stepping, Cycling, New_Sitting). In this case, I want to create a new column (called "Posture"). This new column (as per the example below) should be like:
Posture
<dbl>
2
3
2
1
1
1
3
4
4
4
Here is an example of my data:
> head(graph_pre,30)
# A tibble: 30 × 11
# Groups: Date [1]
Date Time Axis1 Axis2 Axis3 VM Standing Stepping Cycling New_Sitting
<date> <dttm> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 2022-03-14 2022-03-14 09:51:00 89 41 39 105. 0 0 2 0
2 2022-03-14 2022-03-14 09:51:01 88 135 117 199. 0 3 0 0
3 2022-03-14 2022-03-14 09:51:02 0 61 8 61.5 0 0 2 0
4 2022-03-14 2022-03-14 09:51:03 0 25 0 25 0 0 0 1
5 2022-03-14 2022-03-14 09:51:04 0 0 0 0 0 0 0 1
6 2022-03-14 2022-03-14 09:51:05 0 0 0 0 0 0 0 1
7 2022-03-14 2022-03-14 09:51:06 0 24 35 42.4 0 3 0 0
8 2022-03-14 2022-03-14 09:51:07 0 28 0 28 4 0 0 0
9 2022-03-14 2022-03-14 09:51:08 4 96 20 98.1 4 0 0 0
10 2022-03-14 2022-03-14 09:51:09 0 11 0 11 4 0 0 0
# … with 20 more rows, and 1 more variable: Counter <int>
Please let me know if you need more information as I'm new to this.
EDIT
> dput(head(graph_pre,30))
structure(list(Date = structure(c(19065, 19065, 19065, 19065,
19065, 19065, 19065, 19065, 19065, 19065, 19065, 19065, 19065,
19065, 19065, 19065, 19065, 19065, 19065, 19065, 19065, 19065,
19065, 19065, 19065, 19065, 19065, 19065, 19065, 19065), class = "Date"),
Time = structure(c(1647265860, 1647265861, 1647265862, 1647265863,
1647265864, 1647265865, 1647265866, 1647265867, 1647265868,
1647265869, 1647265870, 1647265871, 1647265872, 1647265873,
1647265874, 1647265875, 1647265876, 1647265877, 1647265878,
1647265879, 1647265880, 1647265881, 1647265882, 1647265883,
1647265884, 1647265885, 1647265886, 1647265887, 1647265888,
1647265889), tzone = "", class = c("POSIXct", "POSIXt")),
Axis1 = c(89, 88, 0, 0, 0, 0, 0, 0, 4, 0, 3, 9, 5, 0, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 13, 11, 3, 0), Axis2 = c(41,
135, 61, 25, 0, 0, 24, 28, 96, 11, 91, 44, 8, 8, 29, 1, 17,
0, 0, 0, 15, 0, 0, 0, 0, 28, 47, 28, 48, 0), Axis3 = c(39,
117, 8, 0, 0, 0, 35, 0, 20, 0, 22, 2, 16, 21, 48, 3, 35,
0, 5, 29, 32, 0, 0, 0, 0, 4, 26, 68, 5, 0), VM = c(105.47,
199.14, 61.52, 25, 0, 0, 42.44, 28, 98.14, 11, 93.67, 44.96,
18.57, 22.47, 56.09, 3.16, 38.91, 0, 5, 29, 35.34, 0, 0,
0, 0, 28.28, 55.26, 74.36, 48.35, 0), Standing = c(0, 0,
0, 0, 0, 0, 0, 4, 4, 4, 4, 4, 4, 4, 0, 4, 0, 4, 4, 0, 0,
4, 4, 4, 4, 4, 0, 0, 4, 4), Stepping = c(0, 3, 0, 0, 0, 0,
3, 0, 0, 0, 0, 0, 0, 0, 3, 0, 3, 0, 0, 3, 3, 0, 0, 0, 0,
0, 3, 3, 0, 0), Cycling = c(2, 0, 2, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0), New_Sitting = c(0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0), Counter = c(0L,
1L, 1L, 1L, 0L, 0L, 1L, 1L, 0L, 0L, 0L, 0L, 0L, 0L, 1L, 1L,
1L, 1L, 0L, 1L, 0L, 1L, 0L, 0L, 0L, 0L, 1L, 0L, 1L, 0L)), class = c("grouped_df",
"tbl_df", "tbl", "data.frame"), row.names = c(NA, -30L), groups = structure(list(
Date = structure(19065, class = "Date"), .rows = structure(list(
1:30), ptype = integer(0), class = c("vctrs_list_of",
"vctrs_vctr", "list"))), class = c("tbl_df", "tbl", "data.frame"
), row.names = c(NA, -1L), .drop = TRUE))
What you can do is first replace the zeros with NA and after that unite the columns together. You can use the following code:
library(dplyr)
library(tidyr)
graph_pre %>%
mutate(across(Standing:New_Sitting, na_if, 0)) %>%
unite(Posture, Standing:New_Sitting, na.rm = TRUE, sep = '', remove = T) %>%
mutate(Posture = as.numeric(Posture))
Output:
# A tibble: 30 × 8
# Groups: Date [1]
Date Time Axis1 Axis2 Axis3 VM Posture Counter
<date> <dttm> <dbl> <dbl> <dbl> <dbl> <dbl> <int>
1 2022-03-14 2022-03-14 14:51:00 89 41 39 105. 2 0
2 2022-03-14 2022-03-14 14:51:01 88 135 117 199. 3 1
3 2022-03-14 2022-03-14 14:51:02 0 61 8 61.5 2 1
4 2022-03-14 2022-03-14 14:51:03 0 25 0 25 1 1
5 2022-03-14 2022-03-14 14:51:04 0 0 0 0 1 0
6 2022-03-14 2022-03-14 14:51:05 0 0 0 0 1 0
7 2022-03-14 2022-03-14 14:51:06 0 24 35 42.4 3 1
8 2022-03-14 2022-03-14 14:51:07 0 28 0 28 4 1
9 2022-03-14 2022-03-14 14:51:08 4 96 20 98.1 4 0
10 2022-03-14 2022-03-14 14:51:09 0 11 0 11 4 0
# … with 20 more rows
If you just want to merge them by summing the values for each row, you can do this:
library(tidyverse)
your_dataframe %>%
mutate(Posture = sum(Standing, Stepping, Cycling, New_Sitting))
Which will add an extra column called Posture at the end of your dataframe
Related
I have a dataset that looks like the following:
ID RECN EXSTDAT EXSTPDAT EXONGO EX2LD DOSEA DOSFRM DOSFRQ ADURN STUDYST EXSTDAY EXSTPDAY
<int> <dbl> <date> <date> <dbl> <dbl> <dbl> <chr> <chr> <dbl> <date> <dbl> <dbl>
1 1 1 2022-07-08 2022-07-27 0 0 50 Capsule QD 19 2022-07-08 0 19
2 1 2 2022-07-28 2022-08-14 0 1 50 Capsule QD 17 2022-07-08 20 37
3 2 2 2022-06-09 2022-06-09 0 0 50 Capsule QD 0 2022-06-09 0 0
4 2 1 2022-06-14 2022-08-02 0 0 50 Capsule QD 49 2022-06-09 5 54
5 2 3 2022-08-03 2022-08-14 0 0 0 Capsule QD 11 2022-06-09 55 66
6 2 5 2022-08-15 2022-09-26 0 0 50 Capsule QD 42 2022-06-09 67 109
7 2 4 2022-09-27 2023-02-15 1 0 100 Capsule QD 141 2022-06-09 110 251
8 3 1 2022-06-30 2022-08-03 0 1 50 Capsule QD 34 2022-06-30 0 34
9 4 1 2022-08-24 2022-10-04 0 1 100 Capsule QD 41 2022-08-24 0 41
10 5 1 2022-12-30 2023-01-19 0 1 200 Capsule QD 20 2022-12-30 0 20
I would like to generate an observation for each day between the intervals of EXSTDAY and EXSTPDAY, keeping ID, DOSEA, DOSFRM, and DOSFRQ. Below is an example of the desired results for up to study day (STDAY <= 8) for ID 1 & 2:
ID DOSEA DOSFRM DOSFRQ STDAY
1 50 Capsule QD 0
1 50 Capsule QD 1
1 50 Capsule QD 2
1 50 Capsule QD 3
1 50 Capsule QD 4
1 50 Capsule QD 5
1 50 Capsule QD 6
1 50 Capsule QD 7
1 50 Capsule QD 8
2 50 Capsule QD 0
2 50 Capsule QD 5
2 50 Capsule QD 6
2 50 Capsule QD 7
2 50 Capsule QD 8
I have no idea where to start, so any advice is much appreciated!
dput of original dataset:
structure(list(ID = c(1L, 1L, 2L, 2L, 2L, 2L, 2L, 3L, 4L, 5L,
6L, 7L, 8L, 8L, 9L, 10L, 10L, 10L, 10L, 10L, 11L, 11L, 11L, 12L,
12L, 13L, 14L, 14L, 14L, 14L, 14L, 14L), RECN = c(1, 2, 2, 1,
3, 5, 4, 1, 1, 1, 1, 1, 1, 2, 1, 1, 2, 1, 3, 4, 1, 2, 3, 1, 2,
1, 1, 2, 3, 4, 5, 6), EXSTDAT = structure(c(19181, 19201, 19152,
19157, 19207, 19219, 19262, 19173, 19228, 19356, 19356, 19377,
19303, 19326, 19363, 19216, 19220, 19346, 19362, 19365, 19264,
19277, 19282, 19219, 19226, 19310, 19345, 19351, 19352, 19354,
19355, 19370), class = "Date"), EXSTPDAT = structure(c(19200,
19218, 19152, 19206, 19218, 19261, 19403, 19207, 19269, 19376,
19376, 19403, 19325, 19328, 19383, 19216, 19361, 19366, 19364,
19403, 19275, 19281, 19403, 19225, 19226, 19338, 19350, 19351,
19353, 19354, 19369, 19370), class = "Date"), EXONGO = c(0, 0,
0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0), EX2LD = c(0, 1, 0, 0, 0, 0, 0, 1,
1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0,
0, 0, 0), DOSEA = c(50, 50, 50, 50, 0, 50, 100, 50, 100, 200,
200, 100, 100, 100, 200, 100, 100, 100, 100, 100, 100, 100, 100,
50, 0, 100, 200, 0, 200, 0, 200, 0), DOSFRM = c("Capsule", "Capsule",
"Capsule", "Capsule", "Capsule", "Capsule", "Capsule", "Capsule",
"Capsule", "Capsule", "Capsule", "Capsule", "Capsule", "Tablet",
"Capsule", "Capsule", "Capsule", "Capsule", "Tablet", "Tablet",
"Tablet", "Tablet", "Tablet", "Capsule", "Capsule", "Capsule",
"Capsule", "Capsule", "Capsule", "Capsule", "Capsule", "Capsule"
), DOSFRQ = c("QD", "QD", "QD", "QD", "QD", "QD", "QD", "QD",
"QD", "QD", "QD", "QD", "QD", "QD", "QD", "QD", "QD", "QD", "QD",
"QD", "QD", "QD", "QD", "QD", "QD", "QD", "QD", "QD", "QD", "QD",
"QD", "QD"), ADURN = c(19, 17, 0, 49, 11, 42, 141, 34, 41, 20,
20, 26, 22, 2, 20, 0, 141, 20, 2, 38, 11, 4, 121, 6, 0, 28, 5,
0, 1, 0, 14, 0), STUDYST = structure(c(19181, 19181, 19152, 19152,
19152, 19152, 19152, 19173, 19228, 19356, 19356, 19377, 19303,
19303, 19363, 19216, 19216, 19216, 19216, 19216, 19264, 19264,
19264, 19219, 19219, 19310, 19345, 19345, 19345, 19345, 19345,
19345), class = "Date"), EXSTDAY = c(0, 20, 0, 5, 55, 67, 110,
0, 0, 0, 0, 0, 0, 23, 0, 0, 4, 130, 146, 149, 0, 13, 18, 0, 7,
0, 0, 6, 7, 9, 10, 25), EXSTPDAY = c(19, 37, 0, 54, 66, 109,
251, 34, 41, 20, 20, 26, 22, 25, 20, 0, 145, 150, 148, 187, 11,
17, 139, 6, 7, 28, 5, 6, 8, 9, 24, 25)), row.names = c(NA, -32L
), class = c("tbl_df", "tbl", "data.frame"))
If I understand you correctly, you want one row in your data frame for each day that each subject was participating in your study. If that's the case then you can do:
library(tidyverse)
df %>%
rowwise() %>%
reframe(ID, DOSEA, DOSFRM, DOSFRQ,
STDAY = as.numeric(if(EXSTPDAT - EXSTDAT == 0) EXSTDAT - STUDYST else
seq(EXSTPDAT - EXSTDAT) + EXSTDAT - STUDYST - 1))
#># A tibble: 861 x 5
#> ID DOSEA DOSFRM DOSFRQ STDAY
#> <int> <dbl> <chr> <chr> <dbl>
#> 1 1 50 Capsule QD 0
#> 2 1 50 Capsule QD 1
#> 3 1 50 Capsule QD 2
#> 4 1 50 Capsule QD 3
#> 5 1 50 Capsule QD 4
#> 6 1 50 Capsule QD 5
#> 7 1 50 Capsule QD 6
#> 8 1 50 Capsule QD 7
#> 9 1 50 Capsule QD 8
#>10 1 50 Capsule QD 9
#># ... with 851 more rows
#># i Use `print(n = ...)` to see more rows
If you want to generate a sequence between EXSTDAY and EXSTPDAY one approach could be using map2 from purrr as follows. The final select will indicate which columns you wish to retain in the end.
library(tidyverse)
df %>%
mutate(STDAY = map2(EXSTDAY, EXSTPDAY, seq)) %>%
unnest(STDAY) %>%
select(ID, DOSEA, DOSFRM, DOSFRQ, STDAY)
Output
ID DOSEA DOSFRM DOSFRQ STDAY
<int> <dbl> <chr> <chr> <int>
1 1 50 Capsule QD 0
2 1 50 Capsule QD 1
3 1 50 Capsule QD 2
4 1 50 Capsule QD 3
5 1 50 Capsule QD 4
6 1 50 Capsule QD 5
7 1 50 Capsule QD 6
8 1 50 Capsule QD 7
9 1 50 Capsule QD 8
10 1 50 Capsule QD 9
# … with 877 more rows
So here I what I want, I want to plot 4 columns (Standing, Sitting, Stepping, Cycling) vs Time, and have 1 plot per date. I also want the Y scale to be scaled between 0.5 and 4.5, BUT the Y axis be invisible and a legend saying which color is which.
Here is a sample of my data:
> head(graph_pre,30)
Date Time Axis1 Axis2 Axis3 VM Standing Stepping Cycling New_Sitting Counter
1 2022-05-10 2022-05-10 09:01:00 21 40 2 45.22 0 0 2 0 0
2 2022-05-10 2022-05-10 09:01:01 0 36 1 36.01 0 0 0 1 1
3 2022-05-10 2022-05-10 09:01:02 24 1 0 24.02 0 0 0 1 0
4 2022-05-10 2022-05-10 09:01:03 48 31 4 57.28 0 0 2 0 1
5 2022-05-10 2022-05-10 09:01:04 0 6 0 6.00 0 0 0 1 1
6 2022-05-10 2022-05-10 09:01:05 0 0 0 0.00 0 0 0 1 0
7 2022-05-10 2022-05-10 09:01:06 0 0 0 0.00 0 0 0 1 0
8 2022-05-10 2022-05-10 09:01:07 0 0 0 0.00 0 0 0 1 0
9 2022-05-10 2022-05-10 09:01:08 0 5 2 5.39 0 0 0 1 0
10 2022-05-10 2022-05-10 09:01:09 20 33 3 38.70 0 0 0 1 0
11 2022-05-10 2022-05-10 09:01:10 14 26 29 41.39 0 0 2 0 1
12 2022-05-10 2022-05-10 09:01:11 11 0 4 11.70 0 0 0 1 1
13 2022-05-10 2022-05-10 09:01:12 0 0 0 0.00 0 0 0 1 0
14 2022-05-10 2022-05-10 09:01:13 0 0 0 0.00 0 0 0 1 0
15 2022-05-10 2022-05-10 09:01:14 82 126 113 188.07 0 3 0 0 1
16 2022-05-10 2022-05-10 09:01:15 60 64 47 99.52 0 0 2 0 1
17 2022-05-10 2022-05-10 09:01:16 98 140 236 291.38 0 0 2 0 0
18 2022-05-10 2022-05-10 09:01:17 151 118 221 292.52 0 0 2 0 0
19 2022-05-10 2022-05-10 09:01:18 44 13 99 109.11 0 0 2 0 0
20 2022-05-10 2022-05-10 09:01:19 6 6 53 53.67 0 0 2 0 0
21 2022-05-10 2022-05-10 09:01:20 39 8 65 76.22 0 0 2 0 0
22 2022-05-10 2022-05-10 09:01:21 17 20 57 62.75 0 0 2 0 0
23 2022-05-10 2022-05-10 09:01:22 51 46 269 277.63 0 0 2 0 0
24 2022-05-10 2022-05-10 09:01:23 15 45 82 94.73 0 3 0 0 1
25 2022-05-10 2022-05-10 09:01:24 22 34 4 40.69 0 0 2 0 1
26 2022-05-10 2022-05-10 09:01:25 114 93 41 152.73 0 0 2 0 0
27 2022-05-10 2022-05-10 09:01:26 74 67 92 135.75 0 0 2 0 0
28 2022-05-10 2022-05-10 09:01:27 117 9 40 123.98 0 0 2 0 0
29 2022-05-10 2022-05-10 09:01:28 33 15 0 36.25 0 0 0 1 1
30 2022-05-10 2022-05-10 09:01:29 0 0 0 0.00 0 0 0 1 0
I have the code to separate by date, and to "kinda" plot, but I need it for the 4 columns.
graph_pre <- mutate(graph_pre, day = lubridate::day(Date))
ggplot(graph_pre, aes(x = Time, y = Posture))+
geom_point()+
facet_wrap(~day, scales = "free_x")
dput(head(graph_pre,30))
structure(list(Date = structure(c(19122, 19122, 19122, 19122,
19122, 19122, 19122, 19122, 19122, 19122, 19122, 19122, 19122,
19122, 19122, 19122, 19122, 19122, 19122, 19122, 19122, 19122,
19122, 19122, 19122, 19122, 19122, 19122, 19122, 19122), class = "Date"),
Time = structure(c(1652187660, 1652187661, 1652187662, 1652187663,
1652187664, 1652187665, 1652187666, 1652187667, 1652187668,
1652187669, 1652187670, 1652187671, 1652187672, 1652187673,
1652187674, 1652187675, 1652187676, 1652187677, 1652187678,
1652187679, 1652187680, 1652187681, 1652187682, 1652187683,
1652187684, 1652187685, 1652187686, 1652187687, 1652187688,
1652187689), class = c("POSIXct", "POSIXt"), tzone = ""),
Axis1 = c(21, 0, 24, 48, 0, 0, 0, 0, 0, 20, 14, 11, 0, 0,
82, 60, 98, 151, 44, 6, 39, 17, 51, 15, 22, 114, 74, 117,
33, 0), Axis2 = c(40, 36, 1, 31, 6, 0, 0, 0, 5, 33, 26, 0,
0, 0, 126, 64, 140, 118, 13, 6, 8, 20, 46, 45, 34, 93, 67,
9, 15, 0), Axis3 = c(2, 1, 0, 4, 0, 0, 0, 0, 2, 3, 29, 4,
0, 0, 113, 47, 236, 221, 99, 53, 65, 57, 269, 82, 4, 41,
92, 40, 0, 0), VM = c(45.22, 36.01, 24.02, 57.28, 6, 0, 0,
0, 5.39, 38.7, 41.39, 11.7, 0, 0, 188.07, 99.52, 291.38,
292.52, 109.11, 53.67, 76.22, 62.75, 277.63, 94.73, 40.69,
152.73, 135.75, 123.98, 36.25, 0), Standing = c(0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0), Stepping = c(0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 0,
0, 0, 0, 0), Cycling = c(2, 0, 0, 2, 0, 0, 0, 0, 0, 0, 2,
0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 0, 0),
New_Sitting = c(0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1), Counter = c(0L,
1L, 0L, 1L, 1L, 0L, 0L, 0L, 0L, 0L, 1L, 1L, 0L, 0L, 1L, 1L,
0L, 0L, 0L, 0L, 0L, 0L, 0L, 1L, 1L, 0L, 0L, 0L, 1L, 0L)), row.names = c(NA,
30L), class = "data.frame")
First thing, we should pivot_longer to pull the four posture columns into name-value pairs. Here I've put the names into the "Posture" column. Then we can map that to color and use the values for the y axis.
I've specified the range in scale_y_continuous, but it could also be done with coord_cartesian(ylim = c(0.5,4.5)) -- the difference will be that the out of range points are filtered out in this way, but are in some sense "still there" if you use the coord_cartesian option. That can make a difference if you are doing a summary step, like geom_boxplot or geom_smooth.
Finally, I use theme to specify the y-axis related elements that should be hidden.
library(tidyverse)
graph %>%
mutate(day = lubridate::day(Date)) %>%
pivot_longer(Standing:New_Sitting, names_to = "Posture") %>%
ggplot(aes(x = Time, y = value, color = Posture))+
geom_point()+
scale_y_continuous(limits = c(0.5,4.5), expand = expansion(0)) +
facet_wrap(~day, scales = "free_x") +
labs(title = "Posture vs. Time") +
theme(axis.title.y = element_blank(),
axis.text.y = element_blank(),
axis.ticks.y = element_blank(),
panel.grid.major.y = element_blank(),
panel.grid.minor.y = element_blank())
Here you go:
library(tidyverse)
graph_pre_long <- graph_pre %>% pivot_longer(c(Standing, New_Sitting , Stepping, Cycling), names_to = "Posture")
ggplot(graph_pre_long, aes(x = Time, y = value, color = Posture))+
geom_point()+
facet_wrap(~day, scales = "free_x") +
ylim(.5, 4.5) +
theme(axis.title.y = element_blank(), axis.text.y = element_blank(), axis.ticks.y = element_blank())
require(gtsummary)
test <- structure(list(`1` = c(0, 0, 0, 0, 0, 0, 0, 0, 1, 0), `2` = c(1,0, 0, 0, 0, 1, 0, 1, 0, 0), `3` = c(0, 0, 0, 0, 0, 0, 0, 0, 0,0), `4` = c(1, 1, 0, 0, 1, 0, 0, 0, 0, 0), `5` = c(1, 0, 1, 1,0, 1, 1, 0, 0, 0), `6` = c(0, 0, 0, 1, 0, 0, 1, 0, 0, 0), `7` = c(0,0, 0, 0, 0, 0, 0, 0, 0, 0), `8` = c(0, 0, 0, 0, 0, 0, 0, 0, 0,0), `9` = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0), `10` = c(0, 0, 0,0, 0, 0, 0, 0, 0, 1)), row.names = c(NA, -10L), class = c("tbl_df","tbl", "data.frame"))
In this example data, I have 10 categorical variables.
`1` `2` `3` `4` `5` `6` `7` `8` `9` `10`
<dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 0 1 0 1 1 0 0 0 0 0
2 0 0 0 1 0 0 0 0 0 0
3 0 0 0 0 1 0 0 0 0 0
4 0 0 0 0 1 1 0 0 0 0
5 0 0 0 1 0 0 0 0 0 0
6 0 1 0 0 1 0 0 0 0 0
7 0 0 0 0 1 1 0 0 0 0
8 0 1 0 0 0 0 0 0 0 0
9 1 0 0 0 0 0 0 0 0 0
10 0 0 0 0 0 0 0 0 0 1
Since they can overlap each other, I have put them in different columns,
using 0 and 1, indicatting "yes" or "no" to having (or not having) the categorical variable.
When test %>% tbl_summary(), it creates:
I would like to sort this by frequency, but
test %>% tbl_summary(sort = list(everything() ~ "frequency"))
does not work.
Is there anyway to do this?
Thank you in advance.
The tbl_summary(sort=) argument sorts levels within a variable, not the order the variables appear in the table. Variables are appear in the table in the same order they appear in the data frame.
We can update the order in the data frame using the code below.
library(gtsummary)
#> #Uighur
packageVersion("gtsummary")
#> [1] '1.5.0'
test <- structure(list(`1` = c(0, 0, 0, 0, 0, 0, 0, 0, 1, 0), `2` = c(1,0, 0, 0, 0, 1, 0, 1, 0, 0), `3` = c(0, 0, 0, 0, 0, 0, 0, 0, 0,0), `4` = c(1, 1, 0, 0, 1, 0, 0, 0, 0, 0), `5` = c(1, 0, 1, 1,0, 1, 1, 0, 0, 0), `6` = c(0, 0, 0, 1, 0, 0, 1, 0, 0, 0), `7` = c(0,0, 0, 0, 0, 0, 0, 0, 0, 0), `8` = c(0, 0, 0, 0, 0, 0, 0, 0, 0,0), `9` = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0), `10` = c(0, 0, 0,0, 0, 0, 0, 0, 0, 1)), row.names = c(NA, -10L), class = c("tbl_df","tbl", "data.frame"))
# order variables by prevelence
prev <- purrr::map_dbl(test, mean) %>% sort(decreasing = TRUE)
test %>%
select(all_of(names(prev))) %>%
tbl_summary() %>%
as_kable() # convert to kable for SO
Characteristic
N = 10
5
5 (50%)
2
3 (30%)
4
3 (30%)
6
2 (20%)
1
1 (10%)
10
1 (10%)
3
0 (0%)
7
0 (0%)
8
0 (0%)
9
0 (0%)
Created on 2021-12-10 by the reprex package (v2.0.1)
This question already has answers here:
How to select the rows with maximum values in each group with dplyr? [duplicate]
(6 answers)
Closed 2 years ago.
I want to group by data set based on some IDs, then leave the grouped data that has largest value in the column. Here is a description of my data set.
BSTN ASTN1 BSTN2 ASTN2 BSTN3 ASTN3 BSTN4 ASTN4 BSTN5 ASTN TRNID TRNID2 TRNID3 TRNID4 TRNID5 count
1 150 0 0 0 0 0 0 0 0 152 1674 0 0 0 0 1
2 150 0 0 0 0 0 0 0 0 152 1676 0 0 0 0 2
3 150 0 0 0 0 0 0 0 0 152 1678 0 0 0 0 2
4 150 0 0 0 0 0 0 0 0 152 1680 0 0 0 0 13
5 150 0 0 0 0 0 0 0 0 152 1682 0 0 0 0 3
6 150 0 0 0 0 0 0 0 0 152 1684 0 0 0 0 4
I want to group and summarise this data into a single row based on IDs the first 10 columns BSTN ASTN1 BSTN2 ASTN2 BSTN3 ASTN3 BSTN4 ASTN4 BSTN5 ASTN.
Then for the rest of the columns, TRNID TRNID2 TRNID3 TRNID4 TRNID5 I would like to replace them with the row with maximum value in column count.
What I want as my final output would look as below.
BSTN ASTN1 BSTN2 ASTN2 BSTN3 ASTN3 BSTN4 ASTN4 BSTN5 ASTN TRNID TRNID2 TRNID3 TRNID4 TRNID5 count
150 0 0 0 0 0 0 0 0 152 1680 0 0 0 0 13
How would summarise my data? I have 2,931,959 rows with more groups of BSTN, ASTNs.
dput(head(A_Routetable2))
structure(list(BSTN = c(150, 150, 150, 150, 150, 150), ASTN1 = c(0,
0, 0, 0, 0, 0), BSTN2 = c(0, 0, 0, 0, 0, 0), ASTN2 = c(0, 0,
0, 0, 0, 0), BSTN3 = c(0, 0, 0, 0, 0, 0), ASTN3 = c(0, 0, 0,
0, 0, 0), BSTN4 = c(0, 0, 0, 0, 0, 0), ASTN4 = c(0, 0, 0, 0,
0, 0), BSTN5 = c(0, 0, 0, 0, 0, 0), ASTN = c(152, 152, 152, 152,
152, 152), TRNID = c(1674, 1676, 1678, 1680, 1682, 1684), TRNID2 = c(0,
0, 0, 0, 0, 0), TRNID3 = c(0, 0, 0, 0, 0, 0), TRNID4 = c(0, 0,
0, 0, 0, 0), TRNID5 = c(0, 0, 0, 0, 0, 0), count = c(1L, 2L,
2L, 13L, 3L, 4L)), row.names = c(NA, -6L), groups = structure(list(
BSTN = c(150, 150, 150, 150, 150, 150), ASTN1 = c(0, 0, 0,
0, 0, 0), BSTN2 = c(0, 0, 0, 0, 0, 0), ASTN2 = c(0, 0, 0,
0, 0, 0), BSTN3 = c(0, 0, 0, 0, 0, 0), ASTN3 = c(0, 0, 0,
0, 0, 0), BSTN4 = c(0, 0, 0, 0, 0, 0), ASTN4 = c(0, 0, 0,
0, 0, 0), BSTN5 = c(0, 0, 0, 0, 0, 0), ASTN = c(152, 152,
152, 152, 152, 152), TRNID = c(1674, 1676, 1678, 1680, 1682,
1684), TRNID2 = c(0, 0, 0, 0, 0, 0), TRNID3 = c(0, 0, 0,
0, 0, 0), TRNID4 = c(0, 0, 0, 0, 0, 0), .rows = structure(list(
1L, 2L, 3L, 4L, 5L, 6L), ptype = integer(0), class = c("vctrs_list_of",
"vctrs_vctr", "list"))), row.names = c(NA, 6L), class = c("tbl_df",
"tbl", "data.frame"), .drop = TRUE), class = c("grouped_df",
"tbl_df", "tbl", "data.frame"))
You can group_by position and then select row with max value in count.
library(dplyr)
df %>% group_by(across(1:10)) %>% slice(which.max(count))
# BSTN ASTN1 BSTN2 ASTN2 BSTN3 ASTN3 BSTN4 ASTN4 BSTN5 ASTN TRNID TRNID2 TRNID3 TRNID4 TRNID5 count
# <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int> <int>
#1 150 0 0 0 0 0 0 0 0 152 1680 0 0 0 0 13
Or group by range of columns
df %>% group_by(across(BSTN:ASTN)) %>%slice(which.max(count))
The dput shared by OP is grouped which results an error with across. We can ungroup the data first and run the above which runs without any error. However functions in the previous version of dplyr work without any error on it. For example - group_by_at
A_Routetable2 %>% group_by_at(1:10) %>% slice(which.max(count))
I want to fit a linear regression model using the tsibble package and I have a bunch of dummy variables that I want to include in my analysis. A sample dataset would be the following:
library(tsibble)
library(dplyr)
library(fable)
ex = structure(list(id = c("KEY1", "KEY1", "KEY1", "KEY1", "KEY1",
"KEY1", "KEY1", "KEY1", "KEY1", "KEY1", "KEY1", "KEY1", "KEY1",
"KEY1", "KEY1"), sales = c(0, 5, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0), date = structure(c(15003, 15004, 15005, 15006, 15007,
15008, 15009, 15010, 15011, 15012, 15013, 15014, 15015, 15016,
15017), class = "Date"), wday = c(1L, 2L, 3L, 4L, 5L, 6L, 7L,
1L, 2L, 3L, 4L, 5L, 6L, 7L, 1L), dummy_1 = c(0, 0, 0, 1, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0), dummy_2 = c(0, 0, 0, 0, 0, 0, 1,
0, 0, 0, 0, 0, 0, 0, 0), dummy_3 = c(0, 0, 1, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0)), row.names = c(NA, -15L), key = structure(list(
id = "KEY1", .rows = list(1:15)), row.names = c(NA, -1L), class = c("tbl_df",
"tbl", "data.frame"), .drop = TRUE), index = structure("date", ordered = TRUE), index2 = "date", interval = structure(list(
year = 0, quarter = 0, month = 0, week = 0, day = 1, hour = 0,
minute = 0, second = 0, millisecond = 0, microsecond = 0,
nanosecond = 0, unit = 0), class = "interval"), class = c("tbl_ts",
"tbl_df", "tbl", "data.frame"))
> ex
# A tsibble: 15 x 7 [1D]
# Key: id [1]
id sales date wday dummy_1 dummy_2 dummy_3
<chr> <dbl> <date> <int> <dbl> <dbl> <dbl>
1 KEY1 0 2011-01-29 1 0 0 0
2 KEY1 5 2011-01-30 2 0 0 0
3 KEY1 0 2011-01-31 3 0 0 1
4 KEY1 0 2011-02-01 4 1 0 0
5 KEY1 0 2011-02-02 5 0 0 0
6 KEY1 0 2011-02-03 6 0 0 0
7 KEY1 0 2011-02-04 7 0 1 0
8 KEY1 0 2011-02-05 1 0 0 0
9 KEY1 0 2011-02-06 2 0 0 0
10 KEY1 0 2011-02-07 3 0 0 0
11 KEY1 0 2011-02-08 4 0 0 0
12 KEY1 0 2011-02-09 5 0 0 0
13 KEY1 0 2011-02-10 6 0 0 0
14 KEY1 0 2011-02-11 7 0 0 0
15 KEY1 0 2011-02-12 1 0 0 0
They are too many dummies to specify manually so I was hoping for something faster. Normally I would use the . symbol in the formula in the following way:
fit = ex %>%
model(TSLM(sales ~ trend() + season() + .))
But this does not work:
Warning message:
1 error encountered for TSLM(sales ~ trend() + season() + .)
[1] '.' in formula and no 'data' argument
Is there a systematic tsibble way around this or do I have to create the formula on the fly using the names of the dataset?
We could create a formula with reformulate using the 'dummy' column names
nm1 <- names(ex)[startsWith(names(ex), 'dummy')]
ex %>%
model(lm = TSLM(reformulate(c(nm1, 'trend()', 'season()'), 'sales') ))