Splitting multiple columns using mutate - dplyr
Question
I'm trying to split multiple columns, but struggling to do this efficiently.
I have a df:
> dput(df)
structure(list(UNIQUE_PATIENT_ID = c("DIS-1101-1001-E1", "DIS-1101-1002-E1",
"DIS-1101-1003-E1", "DIS-1101-1004-E1", "DIS-1101-1004-E2", "DIS-1101-1005-E1",
"DIS-1101-1006-E1", "DIS-1101-1007-E1", "DIS-1101-1008-E1", "DIS-1101-1009-E1",
"DIS-1101-1010-E1", "DIS-1101-1011-E1", "DIS-1101-1012-E1", "DIS-1101-1013-E1",
"DIS-1101-1014-E1", "DIS-1101-1015-E1", "DIS-1101-1016-E1", "DIS-1101-1017-E1",
"DIS-1101-1018-E1", "DIS-1101-1019-E1", "DIS-1101-1020-E1", "DIS-1101-1021-E1",
"DIS-1101-1022-E1", "DIS-1101-1023-E1", "DIS-1101-1025-E1", "DIS-1101-1026-E1",
"DIS-1101-1027-E1", "DIS-1101-1028-E1", "DIS-1101-1029-E1", "DIS-1101-1030-E1",
"DIS-1101-1031-E1", "DIS-1101-1032-E1", "DIS-1101-1033-E1", "DIS-1101-1034-E1",
"DIS-1101-1035-E1", "DIS-1101-1036-E1", "DIS-1101-1037-E1", "DIS-1101-1038-E1",
"DIS-1101-1039-E1", "DIS-1101-1040-E1", "DIS-1101-1041-E1", "DIS-1101-1042-E1",
"DIS-1101-1042-E2", "DIS-1101-1043-E1", "DIS-1101-1044-E1", "DIS-1101-1045-E1",
"DIS-1101-1046-E1", "DIS-1101-1047-E1", "DIS-1101-1048-E1", "DIS-1101-1049-E1"
), `TIME_POINT_1=HAEMOGLOBIN` = c("BASELINE=123.00;FIRST=117.00",
"FIRST=92.00", "FIRST=111.00", "BASELINE=125.00;FIRST=113.00",
"FIRST=124.00", "NULL=NULL", "NULL=NULL", "FIRST=77.00", "FIRST=132.00",
"FIRST=123.00", "BASELINE=119.00;FIRST=93.00", "BASELINE=163.00;FIRST=141.00",
"FIRST=137.00", "NULL=NULL", "NULL=NULL", "FIRST=137.00", "FIRST=127.00",
"BASELINE=NULL;FIRST=101.00;SECOND=105.00", "FIRST=120.00", "FIRST=145.00",
"FIRST=126.00", "BASELINE=93.00;FIRST=82.00", "BASELINE=114.00;FIRST=95.00",
"BASELINE=137.00;FIRST=125.00", "BASELINE=157.00;FIRST=112.00",
"NULL=NULL", "FIRST=127.00", "FIRST=147.00", "BASELINE=80.00;FIRST=92.00",
"BASELINE=129.00;FIRST=131.00", "FIRST=108.00", "BASELINE=92.00;FIRST=80.00",
"BASELINE=118.00;FIRST=107.00", "NULL=NULL", "NULL=NULL", "BASELINE=NULL;FIRST=89.00",
"NULL=NULL", "BASELINE=NULL;FIRST=101.00", "FIRST=151.00", "FIRST=106.00",
"BASELINE=81.00;FIRST=87.00", "FIRST=94.00;SECOND=85.00", "FIRST=85.00",
"FIRST=118.00;SECOND=116.00", "FIRST=101.00;SECOND=103.00", "FIRST=114.00",
"BASELINE=NULL;FIRST=129.00", "BASELINE=123.00;FIRST=105.00",
"NULL=NULL", "NULL=NULL"), `TIME_POINT_2=WHITE_CELLS` = c("BASELINE=2.90;FIRST=2.40",
"FIRST=14.80", "FIRST=11.00", "BASELINE=6.90;FIRST=7.30", "FIRST=4.50",
"NULL=NULL", "NULL=NULL", "FIRST=6.10", "FIRST=7.70", "FIRST=16.20",
"BASELINE=7.40;FIRST=5.30", "BASELINE=12.80;FIRST=10.30", "FIRST=14.50",
"NULL=NULL", "NULL=NULL", "FIRST=12.80", "FIRST=3.70", "BASELINE=4.20;FIRST=4.70;SECOND=6.50",
"FIRST=16.60", "FIRST=9.30", "FIRST=15.90", "BASELINE=31.00;FIRST=31.70",
"BASELINE=3.70;FIRST=7.00", "BASELINE=14.10;FIRST=7.20", "BASELINE=21.80;FIRST=31.20",
"NULL=NULL", "FIRST=9.20", "FIRST=5.90", "BASELINE=2.20;FIRST=0.50",
"BASELINE=10.50;FIRST=10.60", "FIRST=12.00", "BASELINE=14.70;FIRST=11.80",
"BASELINE=13.40;FIRST=13.00", "NULL=NULL", "NULL=NULL", "BASELINE=NULL;FIRST=18.60",
"NULL=NULL", "BASELINE=4.30;FIRST=5.90", "FIRST=7.60", "FIRST=3.10",
"BASELINE=10.10;FIRST=11.40", "FIRST=12.00;SECOND=21.30", "FIRST=9.80",
"FIRST=29.50;SECOND=4.80", "FIRST=4.60;SECOND=9.90", "FIRST=18.90",
"BASELINE=15.00;FIRST=13.00", "BASELINE=12.80;FIRST=10.60", "NULL=NULL",
"NULL=NULL")), row.names = c(NA, -50L), class = c("tbl_df", "tbl",
"data.frame"))
I want this output:
dput(output)
structure(list(UNIQUE_PATIENT_ID = c("DIS-1101-1001-E1", "DIS-1101-1002-E1",
"DIS-1101-1003-E1", "DIS-1101-1004-E1", "DIS-1101-1004-E2", "DIS-1101-1005-E1",
"DIS-1101-1006-E1", "DIS-1101-1007-E1", "DIS-1101-1008-E1", "DIS-1101-1009-E1",
"DIS-1101-1010-E1", "DIS-1101-1011-E1", "DIS-1101-1012-E1", "DIS-1101-1013-E1",
"DIS-1101-1014-E1", "DIS-1101-1015-E1", "DIS-1101-1016-E1", "DIS-1101-1017-E1",
"DIS-1101-1018-E1", "DIS-1101-1019-E1", "DIS-1101-1020-E1", "DIS-1101-1021-E1",
"DIS-1101-1022-E1", "DIS-1101-1023-E1", "DIS-1101-1025-E1", "DIS-1101-1026-E1",
"DIS-1101-1027-E1", "DIS-1101-1028-E1", "DIS-1101-1029-E1", "DIS-1101-1030-E1",
"DIS-1101-1031-E1", "DIS-1101-1032-E1", "DIS-1101-1033-E1", "DIS-1101-1034-E1",
"DIS-1101-1035-E1", "DIS-1101-1036-E1", "DIS-1101-1037-E1", "DIS-1101-1038-E1",
"DIS-1101-1039-E1", "DIS-1101-1040-E1", "DIS-1101-1041-E1", "DIS-1101-1042-E1",
"DIS-1101-1042-E2", "DIS-1101-1043-E1", "DIS-1101-1044-E1", "DIS-1101-1045-E1",
"DIS-1101-1046-E1", "DIS-1101-1047-E1", "DIS-1101-1048-E1", "DIS-1101-1049-E1"
), Haemoglobin_BASELINE = c("123.00", NA, NA, "125.00", NA, NA,
NA, NA, NA, NA, "119.00", "163.00", NA, NA, NA, NA, NA, "NULL",
NA, NA, NA, "93.00", "114.00", "137.00", "157.00", NA, NA, NA,
"80.00", "129.00", NA, "92.00", "118.00", NA, NA, "NULL", NA,
"NULL", NA, NA, "81.00", NA, NA, NA, NA, NA, "NULL", "123.00",
NA, NA), Haemoglobin_FIRST = c("117.00", "92.00", "111.00", "113.00",
"124.00", NA, NA, "77.00", "132.00", "123.00", "93.00", "141.00",
"137.00", NA, NA, "137.00", "127.00", "101.00", "120.00", "145.00",
"126.00", "82.00", "95.00", "125.00", "112.00", NA, "127.00",
"147.00", "92.00", "131.00", "108.00", "80.00", "107.00", NA,
NA, "89.00", NA, "101.00", "151.00", "106.00", "87.00", "94.00",
"85.00", "118.00", "101.00", "114.00", "129.00", "105.00", NA,
NA), Haemoglobin_SECOND = c(NA, NA, NA, NA, NA, NA, NA, NA, NA,
NA, NA, NA, NA, NA, NA, NA, NA, "105.00", NA, NA, NA, NA, NA,
NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,
NA, NA, "85.00", NA, "116.00", "103.00", NA, NA, NA, NA, NA),
WBC_BASELINE = c("2.90", NA, NA, "6.90", NA, NA, NA, NA,
NA, NA, "7.40", "12.80", NA, NA, NA, NA, NA, "4.20", NA,
NA, NA, "31.00", "3.70", "14.10", "21.80", NA, NA, NA, "2.20",
"10.50", NA, "14.70", "13.40", NA, NA, "NULL", NA, "4.30",
NA, NA, "10.10", NA, NA, NA, NA, NA, "15.00", "12.80", NA,
NA), WBC_FIRST = c("2.40", "14.80", "11.00", "7.30", "4.50",
NA, NA, "6.10", "7.70", "16.20", "5.30", "10.30", "14.50",
NA, NA, "12.80", "3.70", "4.70", "16.60", "9.30", "15.90",
"31.70", "7.00", "7.20", "31.20", NA, "9.20", "5.90", "0.50",
"10.60", "12.00", "11.80", "13.00", NA, NA, "18.60", NA,
"5.90", "7.60", "3.10", "11.40", "12.00", "9.80", "29.50",
"4.60", "18.90", "13.00", "10.60", NA, NA), WBC_SECOND = c(NA,
NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,
NA, "6.50", NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,
NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, "21.30", NA,
"4.80", "9.90", NA, NA, NA, NA, NA)), row.names = c(NA, -50L
), class = "data.frame")
I can achieve this, but my code is inefficient (as I am doing this on each blood column):
#haemoglobin
haem_ID <- df %>% select(UNIQUE_PATIENT_ID, `TIME_POINT_1=HAEMOGLOBIN`)
haem_sep <- haem_ID %>%
rename(text = `TIME_POINT_1=HAEMOGLOBIN`) %>%
separate_rows(text, sep = ';') %>%
separate(text, c('text', 'value'), sep = '=') %>%
pivot_wider(names_from = text, values_from = value)
tmp <- haem_sep %>% select(-UNIQUE_PATIENT_ID)
colnames(tmp) <- paste("Haemoglobin", colnames(tmp), sep = "_")
tmp2 <- haem_ID %>% select(UNIQUE_PATIENT_ID)
cbind(tmp2, tmp) -> tmp3
tmp3 %<>% select(-Haemoglobin_NULL)
haem <- tmp3
#white cells
a <- df %>% select(UNIQUE_PATIENT_ID, `TIME_POINT_2=WHITE_CELLS`)
b <- a %>%
rename(text = `TIME_POINT_2=WHITE_CELLS`) %>%
separate_rows(text, sep = ';') %>%
separate(text, c('text', 'value'), sep = '=') %>%
pivot_wider(names_from = text, values_from = value)
tmp <- b %>% select(-UNIQUE_PATIENT_ID)
colnames(tmp) <- paste("WBC", colnames(tmp), sep = "_")
tmp2 <- a %>% select(UNIQUE_PATIENT_ID)
cbind(tmp2, tmp) -> tmp3
tmp3 %<>% select(-WBC_NULL)
WBC <- tmp3
output <- merge(haem, WBC)
Pls can someone help me achieve the same but more effectively please? In reality I have 20 blood tests and not just 2, so hard-coding it is not ideal.
UPDATE
Upon the kind responses, I've tried running the suggested code but I'm getting an error and I can't figure out why.
input data
structure(list(UNIQUE_PATIENT_ID = c("DIS-1104-1021-E1", "DIS-1107-1010-E1",
"DIS-1124-1108-E1"), `TIME_POINT_1=HAEMOGLOBIN` = c("BASELINE=98.00;FIRST=90.00;SECOND=81.00;FIRST=85.00",
"FIRST=120.00;NULL=120.00;NULL=118.00;SECOND=109.00;NULL=186.00",
"BASELINE=78.00;FIRST=NULL;FIRST=99.00"), `TIME_POINT_2=WHITE_CELLS` = c("BASELINE=0.01;FIRST=NULL;SECOND=0.01;FIRST=0.00",
"FIRST=15.40;NULL=16.10;NULL=14.50;SECOND=11.40;NULL=2.20;NULL=5.00",
"BASELINE=85.00;FIRST=NULL;FIRST=209.00"), `TIME_POINT_3=PLATELETS` = c("BASELINE=28.00;FIRST=28.00;SECOND=12.00;FIRST=21.00",
"FIRST=200.00;NULL=186.00;NULL=203.00;SECOND=243.00;NULL=145.00;NULL=457.00",
"BASELINE=186.00;FIRST=NULL"), `TIME_POINT_4=NEUTROPHILS` = c("BASELINE=0.00;FIRST=0.00;SECOND=0.00",
"FIRST=13.00;NULL=13.20;NULL=11.10;SECOND=7.70;NULL=1.30;NULL=2.20",
"BASELINE=7.10;FIRST=2.40;FIRST=NULL"), `TIME_POINT_5=LYMPHOCYTES` = c("BASELINE=0.00;FIRST=0.00;SECOND=0.00",
"FIRST=0.80;NULL=1.60;NULL=1.40;SECOND=1.70;NULL=0.40", "BASELINE=76.80;FIRST=135.00;FIRST=NULL"
), `TIME_POINT_6=MONOCYTES` = c("BASELINE=0.00;FIRST=0.00;SECOND=0.00",
"FIRST=1.50;NULL=1.20;NULL=1.90;SECOND=2.00;NULL=0.40;NULL=0.90",
"BASELINE=1.40;FIRST=4.20;FIRST=NULL"), `TIME_POINT_7=EOSINOPHILS` = c("BASELINE=0.00;FIRST=0.00;SECOND=0.00",
"FIRST=0.10;NULL=0.00;SECOND=0.00;NULL=0.10;NULL=0.30", "BASELINE=0.10;FIRST=0.40;FIRST=NULL"
), `TIME_POINT_8=FIBRINOGEN` = c("BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=6.90;NULL=6.60;NULL=7.10;SECOND=6.60;NULL=6.00;NULL=NULL",
"BASELINE=6.70;FIRST=NULL"), `TIME_POINT_9=PT` = c("BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=19.20;NULL=17.20;NULL=16.10;SECOND=14.40;NULL=14.90;NULL=NULL",
"BASELINE=13.40;FIRST=NULL"), `TIME_POINT_10=UREA` = c("BASELINE=4.30;FIRST=4.30;SECOND=5.10;FIRST=4.50",
"FIRST=3.80;NULL=3.70;NULL=4.40;SECOND=2.80;NULL=2.90;NULL=4.10",
"BASELINE=5.10;FIRST=NULL;FIRST=7.20"), `TIME_POINT_11=CREATININE` = c("BASELINE=67.00;FIRST=55.00;SECOND=61.00;FIRST=60.00",
"FIRST=44.00;NULL=42.00;NULL=45.00;SECOND=42.00;NULL=33.00",
"BASELINE=65.00;FIRST=NULL;FIRST=72.00"), `TIME_POINT_12=ALT` = c("BASELINE=9.00;FIRST=11.00;SECOND=8.00;FIRST=5.00",
"FIRST=20.00;NULL=NULL;NULL=19.00;SECOND=16.00;NULL=16.00;NULL=21.00",
"BASELINE=64.00;FIRST=NULL;FIRST=92.00"), `TIME_POINT_13=BILIRUBIN` = c("BASELINE=3.00;FIRST=5.00;SECOND=3.00;FIRST=NULL",
"FIRST=20.00;NULL=NULL;NULL=11.00;SECOND=10.00;NULL=5.00", "BASELINE=7.00;FIRST=NULL;FIRST=10.00"
), `TIME_POINT_14=ALBUMIN` = c("BASELINE=42.00;FIRST=39.00;SECOND=35.00;FIRST=36.00",
"FIRST=42.00;NULL=NULL;NULL=37.00;SECOND=35.00;NULL=35.00;NULL=39.00",
"BASELINE=31.00;FIRST=NULL;FIRST=40.00"), `TIME_POINT_15=VITAMIN_D` = c("BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;SECOND=NULL", "BASELINE=NULL;FIRST=NULL"
), `TIME_POINT_16=CRP` = c("BASELINE=NULL;FIRST=172.00;SECOND=116.00;FIRST=212.00",
"FIRST=427.00;NULL=420.00;NULL=338.00;SECOND=215.00;NULL=87.00;NULL=64.00",
"BASELINE=199.00;FIRST=NULL;FIRST=1.00"), `TIME_POINT_17=PCT` = c("BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;SECOND=NULL", "BASELINE=NULL;FIRST=NULL"
), `TIME_POINT_18=FERRITIN` = c("BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;NULL=378.00;SECOND=NULL", "BASELINE=809.00;FIRST=432.00;FIRST=NULL"
), `TIME_POINT_19=TROP_T` = c("BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;NULL=3.00;SECOND=NULL", "BASELINE=NULL;FIRST=NULL"
), `TIME_POINT_20=D_DIMER` = c("BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=1.85;NULL=1.65;SECOND=NULL;NULL=NULL", "BASELINE=749.00;FIRST=486.00;FIRST=NULL"
), `TIME_POINT_21=INR` = c("BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;SECOND=NULL", "BASELINE=NULL;FIRST=NULL"
), `TIME_POINT_22=BNP_NT` = c("BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;SECOND=NULL", "BASELINE=NULL;FIRST=NULL"
), `TIME_POINT_23=LDH` = c("BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;SECOND=NULL", "BASELINE=NULL;FIRST=254.00;FIRST=NULL"
), `TIME_POINT_24=CK` = c("BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;NULL=29.00;SECOND=NULL", "BASELINE=NULL;FIRST=NULL"
), `TIME_POINT_25=SCD_25` = c("BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;SECOND=NULL", "BASELINE=NULL;FIRST=NULL"
), `TIME_POINT_26=BASE_EXCESS` = c("BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=3.10;NULL=1.50;NULL=NULL;SECOND=NULL", "BASELINE=-3.80;FIRST=NULL"
), `TIME_POINT_27=LACTATE` = c("BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=0.90;NULL=1.30;NULL=NULL;SECOND=NULL", "BASELINE=0.60;FIRST=NULL"
)), row.names = c(NA, -3L), class = c("tbl_df", "tbl", "data.frame"
))
Code:
output <- input |>
separate_rows(-UNIQUE_PATIENT_ID, sep = ";") |>
pivot_longer(-UNIQUE_PATIENT_ID) |>
separate(value, into = c("timepoint", "value"), sep = "=") |>
mutate(name = str_remove(name, ".*=")) |>
pivot_wider(names_from = c("name", "timepoint")) |>
select(-ends_with("NULL"))
Error message:
Error in `fn()`:
! In row 2, can't recycle input of size 5 to size 6.
Run `rlang::last_error()` to see where the error occurred.
Thank you!
2 answers
solution1
1 ACCPTED 2022-07-04 10:07:44
The last select line is optional depending on how you want the columns ordered and whether or not you wish to keep the NULL columns:
df |>
separate_rows(-UNIQUE_PATIENT_ID, sep = ";") |>
pivot_longer(-UNIQUE_PATIENT_ID) |>
separate(value, into = c("timepoint", "value"), sep = "=") |>
mutate(name = str_remove(name, ".*=")) |>
pivot_wider(names_from = c("name", "timepoint")) |>
select(UNIQUE_PATIENT_ID, sort(tidyselect::peek_vars()), -contains("NULL"))
Using the example data with syntactic columns names, ie using a "." instead of "="), the following reprex may be run in a clean R session:
library(tidyverse)
df <- tribble(
~UNIQUE_PATIENT_ID, ~TIME_POINT_1.HAEMOGLOBIN, ~TIME_POINT_2.WHITE_CELLS,
"DIS-1101-1001-E1", "BASELINE=123.00;FIRST=117.00", "BASELINE=2.90;FIRST=2.40",
"DIS-1101-1002-E1", "FIRST=92.00", "FIRST=14.80",
"DIS-1101-1003-E1", "FIRST=111.00", "FIRST=11.00",
"DIS-1101-1004-E1", "BASELINE=125.00;FIRST=113.00", "BASELINE=6.90;FIRST=7.30",
"DIS-1101-1004-E2", "FIRST=124.00", "FIRST=4.50",
"DIS-1101-1005-E1", "NULL=NULL", "NULL=NULL",
"DIS-1101-1006-E1", "NULL=NULL", "NULL=NULL",
"DIS-1101-1007-E1", "FIRST=77.00", "FIRST=6.10",
"DIS-1101-1008-E1", "FIRST=132.00", "FIRST=7.70",
"DIS-1101-1009-E1", "FIRST=123.00", "FIRST=16.20",
"DIS-1101-1010-E1", "BASELINE=119.00;FIRST=93.00", "BASELINE=7.40;FIRST=5.30",
"DIS-1101-1011-E1", "BASELINE=163.00;FIRST=141.00", "BASELINE=12.80;FIRST=10.30",
"DIS-1101-1012-E1", "FIRST=137.00", "FIRST=14.50",
"DIS-1101-1013-E1", "NULL=NULL", "NULL=NULL",
"DIS-1101-1014-E1", "NULL=NULL", "NULL=NULL",
"DIS-1101-1015-E1", "FIRST=137.00", "FIRST=12.80",
"DIS-1101-1016-E1", "FIRST=127.00", "FIRST=3.70",
"DIS-1101-1017-E1", "BASELINE=NULL;FIRST=101.00;SECOND=105.00", "BASELINE=4.20;FIRST=4.70;SECOND=6.50",
"DIS-1101-1018-E1", "FIRST=120.00", "FIRST=16.60",
"DIS-1101-1019-E1", "FIRST=145.00", "FIRST=9.30",
"DIS-1101-1020-E1", "FIRST=126.00", "FIRST=15.90",
"DIS-1101-1021-E1", "BASELINE=93.00;FIRST=82.00", "BASELINE=31.00;FIRST=31.70",
"DIS-1101-1022-E1", "BASELINE=114.00;FIRST=95.00", "BASELINE=3.70;FIRST=7.00",
"DIS-1101-1023-E1", "BASELINE=137.00;FIRST=125.00", "BASELINE=14.10;FIRST=7.20",
"DIS-1101-1025-E1", "BASELINE=157.00;FIRST=112.00", "BASELINE=21.80;FIRST=31.20",
"DIS-1101-1026-E1", "NULL=NULL", "NULL=NULL",
"DIS-1101-1027-E1", "FIRST=127.00", "FIRST=9.20",
"DIS-1101-1028-E1", "FIRST=147.00", "FIRST=5.90",
"DIS-1101-1029-E1", "BASELINE=80.00;FIRST=92.00", "BASELINE=2.20;FIRST=0.50",
"DIS-1101-1030-E1", "BASELINE=129.00;FIRST=131.00", "BASELINE=10.50;FIRST=10.60",
"DIS-1101-1031-E1", "FIRST=108.00", "FIRST=12.00",
"DIS-1101-1032-E1", "BASELINE=92.00;FIRST=80.00", "BASELINE=14.70;FIRST=11.80",
"DIS-1101-1033-E1", "BASELINE=118.00;FIRST=107.00", "BASELINE=13.40;FIRST=13.00",
"DIS-1101-1034-E1", "NULL=NULL", "NULL=NULL",
"DIS-1101-1035-E1", "NULL=NULL", "NULL=NULL",
"DIS-1101-1036-E1", "BASELINE=NULL;FIRST=89.00", "BASELINE=NULL;FIRST=18.60",
"DIS-1101-1037-E1", "NULL=NULL", "NULL=NULL",
"DIS-1101-1038-E1", "BASELINE=NULL;FIRST=101.00", "BASELINE=4.30;FIRST=5.90",
"DIS-1101-1039-E1", "FIRST=151.00", "FIRST=7.60",
"DIS-1101-1040-E1", "FIRST=106.00", "FIRST=3.10",
"DIS-1101-1041-E1", "BASELINE=81.00;FIRST=87.00", "BASELINE=10.10;FIRST=11.40",
"DIS-1101-1042-E1", "FIRST=94.00;SECOND=85.00", "FIRST=12.00;SECOND=21.30",
"DIS-1101-1042-E2", "FIRST=85.00", "FIRST=9.80",
"DIS-1101-1043-E1", "FIRST=118.00;SECOND=116.00", "FIRST=29.50;SECOND=4.80",
"DIS-1101-1044-E1", "FIRST=101.00;SECOND=103.00", "FIRST=4.60;SECOND=9.90",
"DIS-1101-1045-E1", "FIRST=114.00", "FIRST=18.90",
"DIS-1101-1046-E1", "BASELINE=NULL;FIRST=129.00", "BASELINE=15.00;FIRST=13.00",
"DIS-1101-1047-E1", "BASELINE=123.00;FIRST=105.00", "BASELINE=12.80;FIRST=10.60",
"DIS-1101-1048-E1", "NULL=NULL", "NULL=NULL",
"DIS-1101-1049-E1", "NULL=NULL", "NULL=NULL"
)
df |>
separate_rows(-UNIQUE_PATIENT_ID, sep = ";") |>
pivot_longer(-UNIQUE_PATIENT_ID) |>
separate(value, into = c("timepoint", "value"), sep = "=") |>
mutate(name = str_remove(name, ".*\\.")) |>
pivot_wider(names_from = c("name", "timepoint")) |>
select(UNIQUE_PATIENT_ID, sort(tidyselect::peek_vars()), -contains("NULL"))
#> # A tibble: 50 × 7
#> UNIQUE_PATIENT_ID HAEMOGLOBIN_BASELINE HAEMOGLOBIN_FIRST HAEMOGLOBIN_SECOND
#> <chr> <chr> <chr> <chr>
#> 1 DIS-1101-1001-E1 123.00 117.00 <NA>
#> 2 DIS-1101-1002-E1 <NA> 92.00 <NA>
#> 3 DIS-1101-1003-E1 <NA> 111.00 <NA>
#> 4 DIS-1101-1004-E1 125.00 113.00 <NA>
#> 5 DIS-1101-1004-E2 <NA> 124.00 <NA>
#> 6 DIS-1101-1005-E1 <NA> <NA> <NA>
#> 7 DIS-1101-1006-E1 <NA> <NA> <NA>
#> 8 DIS-1101-1007-E1 <NA> 77.00 <NA>
#> 9 DIS-1101-1008-E1 <NA> 132.00 <NA>
#> 10 DIS-1101-1009-E1 <NA> 123.00 <NA>
#> # … with 40 more rows, and 3 more variables: WHITE_CELLS_BASELINE <chr>,
#> # WHITE_CELLS_FIRST <chr>, WHITE_CELLS_SECOND <chr>
Created on 2022-07-04 by the reprex package (v2.0.1)
** UPDATE FOR YOUR WIDER SAMPLE2 DATASET **
To handle the varying number of separators across columns I've suggested a function to separate_rows iteratively by column. This overcomes the separate_rows error in the wider dataset that wasn't an issue in the sample dataset.
Using pattern matching, I've also extracted only strings referencing "BASELINE", "FIRST" or "SECOND" where these are associated with a value.
And as @harre notes, because some columns contain multiple, eg FIRSTs, I've subsequently taken the first of these using slice_head before using pivot_wider . You could instead take the last using slice_tail or take the mean etc. depending on the requirement.
You'll need to test further on your extended dataset, but hope this helps!
library(tidyverse)
df <- structure(list(UNIQUE_PATIENT_ID = c(
"DIS-1104-1021-E1", "DIS-1107-1010-E1",
"DIS-1124-1108-E1"
), `TIME_POINT_1=HAEMOGLOBIN` = c(
"BASELINE=98.00;FIRST=90.00;SECOND=81.00;FIRST=85.00",
"FIRST=120.00;NULL=120.00;NULL=118.00;SECOND=109.00;NULL=186.00",
"BASELINE=78.00;FIRST=NULL;FIRST=99.00"
), `TIME_POINT_2=WHITE_CELLS` = c(
"BASELINE=0.01;FIRST=NULL;SECOND=0.01;FIRST=0.00",
"FIRST=15.40;NULL=16.10;NULL=14.50;SECOND=11.40;NULL=2.20;NULL=5.00",
"BASELINE=85.00;FIRST=NULL;FIRST=209.00"
), `TIME_POINT_3=PLATELETS` = c(
"BASELINE=28.00;FIRST=28.00;SECOND=12.00;FIRST=21.00",
"FIRST=200.00;NULL=186.00;NULL=203.00;SECOND=243.00;NULL=145.00;NULL=457.00",
"BASELINE=186.00;FIRST=NULL"
), `TIME_POINT_4=NEUTROPHILS` = c(
"BASELINE=0.00;FIRST=0.00;SECOND=0.00",
"FIRST=13.00;NULL=13.20;NULL=11.10;SECOND=7.70;NULL=1.30;NULL=2.20",
"BASELINE=7.10;FIRST=2.40;FIRST=NULL"
), `TIME_POINT_5=LYMPHOCYTES` = c(
"BASELINE=0.00;FIRST=0.00;SECOND=0.00",
"FIRST=0.80;NULL=1.60;NULL=1.40;SECOND=1.70;NULL=0.40", "BASELINE=76.80;FIRST=135.00;FIRST=NULL"
), `TIME_POINT_6=MONOCYTES` = c(
"BASELINE=0.00;FIRST=0.00;SECOND=0.00",
"FIRST=1.50;NULL=1.20;NULL=1.90;SECOND=2.00;NULL=0.40;NULL=0.90",
"BASELINE=1.40;FIRST=4.20;FIRST=NULL"
), `TIME_POINT_7=EOSINOPHILS` = c(
"BASELINE=0.00;FIRST=0.00;SECOND=0.00",
"FIRST=0.10;NULL=0.00;SECOND=0.00;NULL=0.10;NULL=0.30", "BASELINE=0.10;FIRST=0.40;FIRST=NULL"
), `TIME_POINT_8=FIBRINOGEN` = c(
"BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=6.90;NULL=6.60;NULL=7.10;SECOND=6.60;NULL=6.00;NULL=NULL",
"BASELINE=6.70;FIRST=NULL"
), `TIME_POINT_9=PT` = c(
"BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=19.20;NULL=17.20;NULL=16.10;SECOND=14.40;NULL=14.90;NULL=NULL",
"BASELINE=13.40;FIRST=NULL"
), `TIME_POINT_10=UREA` = c(
"BASELINE=4.30;FIRST=4.30;SECOND=5.10;FIRST=4.50",
"FIRST=3.80;NULL=3.70;NULL=4.40;SECOND=2.80;NULL=2.90;NULL=4.10",
"BASELINE=5.10;FIRST=NULL;FIRST=7.20"
), `TIME_POINT_11=CREATININE` = c(
"BASELINE=67.00;FIRST=55.00;SECOND=61.00;FIRST=60.00",
"FIRST=44.00;NULL=42.00;NULL=45.00;SECOND=42.00;NULL=33.00",
"BASELINE=65.00;FIRST=NULL;FIRST=72.00"
), `TIME_POINT_12=ALT` = c(
"BASELINE=9.00;FIRST=11.00;SECOND=8.00;FIRST=5.00",
"FIRST=20.00;NULL=NULL;NULL=19.00;SECOND=16.00;NULL=16.00;NULL=21.00",
"BASELINE=64.00;FIRST=NULL;FIRST=92.00"
), `TIME_POINT_13=BILIRUBIN` = c(
"BASELINE=3.00;FIRST=5.00;SECOND=3.00;FIRST=NULL",
"FIRST=20.00;NULL=NULL;NULL=11.00;SECOND=10.00;NULL=5.00", "BASELINE=7.00;FIRST=NULL;FIRST=10.00"
), `TIME_POINT_14=ALBUMIN` = c(
"BASELINE=42.00;FIRST=39.00;SECOND=35.00;FIRST=36.00",
"FIRST=42.00;NULL=NULL;NULL=37.00;SECOND=35.00;NULL=35.00;NULL=39.00",
"BASELINE=31.00;FIRST=NULL;FIRST=40.00"
), `TIME_POINT_15=VITAMIN_D` = c(
"BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;SECOND=NULL", "BASELINE=NULL;FIRST=NULL"
), `TIME_POINT_16=CRP` = c(
"BASELINE=NULL;FIRST=172.00;SECOND=116.00;FIRST=212.00",
"FIRST=427.00;NULL=420.00;NULL=338.00;SECOND=215.00;NULL=87.00;NULL=64.00",
"BASELINE=199.00;FIRST=NULL;FIRST=1.00"
), `TIME_POINT_17=PCT` = c(
"BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;SECOND=NULL", "BASELINE=NULL;FIRST=NULL"
), `TIME_POINT_18=FERRITIN` = c(
"BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;NULL=378.00;SECOND=NULL", "BASELINE=809.00;FIRST=432.00;FIRST=NULL"
), `TIME_POINT_19=TROP_T` = c(
"BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;NULL=3.00;SECOND=NULL", "BASELINE=NULL;FIRST=NULL"
), `TIME_POINT_20=D_DIMER` = c(
"BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=1.85;NULL=1.65;SECOND=NULL;NULL=NULL", "BASELINE=749.00;FIRST=486.00;FIRST=NULL"
), `TIME_POINT_21=INR` = c(
"BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;SECOND=NULL", "BASELINE=NULL;FIRST=NULL"
), `TIME_POINT_22=BNP_NT` = c(
"BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;SECOND=NULL", "BASELINE=NULL;FIRST=NULL"
), `TIME_POINT_23=LDH` = c(
"BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;SECOND=NULL", "BASELINE=NULL;FIRST=254.00;FIRST=NULL"
), `TIME_POINT_24=CK` = c(
"BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;NULL=29.00;SECOND=NULL", "BASELINE=NULL;FIRST=NULL"
), `TIME_POINT_25=SCD_25` = c(
"BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=NULL;NULL=NULL;SECOND=NULL", "BASELINE=NULL;FIRST=NULL"
), `TIME_POINT_26=BASE_EXCESS` = c(
"BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=3.10;NULL=1.50;NULL=NULL;SECOND=NULL", "BASELINE=-3.80;FIRST=NULL"
), `TIME_POINT_27=LACTATE` = c(
"BASELINE=NULL;FIRST=NULL;SECOND=NULL",
"FIRST=0.90;NULL=1.30;NULL=NULL;SECOND=NULL", "BASELINE=0.60;FIRST=NULL"
)), row.names = c(NA, -3L), class = c("tbl_df", "tbl", "data.frame"))
pattern <- c("FIRST=\\d{2,3}.\\d{0,2}|SECOND=\\d{2,3}.\\d{0,2}|BASELINE=\\d{2,3}.\\d{0,2}")
# Function to iteratively separate by column
separate_iteratively <- function(col){
separate_rows(df2, col, sep = ";") |>
select(1, col)
}
# Extract only if BASELINE, FIRST or SECOND with an associate number
df2 <- df |>
mutate(across(-UNIQUE_PATIENT_ID, ~ str_extract_all(., pattern) |>
map_chr(~str_c(., collapse = ";"))))
# Separate & rearrange
df3 <- map_dfr(2:ncol(df2), separate_iteratively) |>
pivot_longer(-UNIQUE_PATIENT_ID) |>
filter(!is.na(value) & !value == "") |>
separate(value, into = c("timepoint", "value"), sep = "=") |>
mutate(name = str_remove(name, ".*=")) |>
arrange(UNIQUE_PATIENT_ID, name, timepoint) |>
group_by(UNIQUE_PATIENT_ID, name, timepoint) |>
slice_head(n = 1) |> # Or use slice_tail
pivot_wider(names_from = c("name", "timepoint"))|>
select(UNIQUE_PATIENT_ID, sort(tidyselect::peek_vars()))
df3
#> # A tibble: 3 × 35
#> # Groups: UNIQUE_PATIENT_ID [3]
#> UNIQUE_PATIENT_ID ALBUMIN_BASELINE ALBUMIN_FIRST ALBUMIN_SECOND ALT_BASELINE
#> <chr> <chr> <chr> <chr> <chr>
#> 1 DIS-1104-1021-E1 42.00 39.00 35.00 <NA>
#> 2 DIS-1107-1010-E1 <NA> 42.00 35.00 <NA>
#> 3 DIS-1124-1108-E1 31.00 40.00 <NA> 64.00
#> # … with 30 more variables: ALT_FIRST <chr>, ALT_SECOND <chr>,
#> # BILIRUBIN_FIRST <chr>, BILIRUBIN_SECOND <chr>, CREATININE_BASELINE <chr>,
#> # CREATININE_FIRST <chr>, CREATININE_SECOND <chr>, CRP_BASELINE <chr>,
#> # CRP_FIRST <chr>, CRP_SECOND <chr>, D_DIMER_BASELINE <chr>,
#> # D_DIMER_FIRST <chr>, FERRITIN_BASELINE <chr>, FERRITIN_FIRST <chr>,
#> # HAEMOGLOBIN_BASELINE <chr>, HAEMOGLOBIN_FIRST <chr>,
#> # HAEMOGLOBIN_SECOND <chr>, LDH_FIRST <chr>, LYMPHOCYTES_BASELINE <chr>, …
# Just to illustrate the issue with varying separators
tribble(~v1, ~v2,
"a;a", "b;b",
"a;a", "b;b;b",
) |>
separate_rows(everything(), sep = ";")
#> Error in `fn()`:
#> ! In row 2, can't recycle input of size 2 to size 3.
# Works
tribble(~v1, ~v2,
"a;a", "b;b",
"a;a", "b;b",
) |>
separate_rows(everything(), sep = ";")
#> # A tibble: 4 × 2
#> v1 v2
#> <chr> <chr>
#> 1 a b
#> 2 a b
#> 3 a b
#> 4 a b
Created on 2022-07-04 by the reprex package (v2.0.1)
solution2
1 2022-07-04 15:48:06
Building on my comments above and @Carl's implementation, you you could utilize the nested list structure directly using values_fn , which might save you from some lines of code. I am using first() from dplyr to take the first element, but you could as well use other functions depending on what you prefer, ~ mean(as.numeric(.)) to take the mean etc. There is no varying number of separators when we pivot_longer before using separate_rows , when I run the data.
df |>
pivot_longer(-UNIQUE_PATIENT_ID) |>
separate_rows(-UNIQUE_PATIENT_ID, sep = ";") |>
separate(value, into = c("timepoint", "value"), sep = "=") |>
mutate(name = str_remove(name, ".*=")) |>
filter(!is.na(value) & value != "NULL" & timepoint != "NULL") |>
pivot_wider(names_from = c("name", "timepoint"), values_fn = ~ first(.)) |>
select(UNIQUE_PATIENT_ID, sort(tidyselect::peek_vars()))
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.
Related Question
Mutate multiple columns with conditions using dplyr
Using dplyr mutate with conditions based on multiple columns
R: Using dplyr to Mutate Multiple Columns
Return multiple columns in dplyr mutate
Mutate multiple / consecutive columns (with dplyr)
dplyr mutate using variable columns
Dplyr mutate using existing columns
Apply Multiple Columns to Custom function Using dplyr::mutate(across())
Using functions of multiple columns in a dplyr mutate_at call
Using dplyr and mutate with separate conditions for each of multiple columns
Related Tags