I have a huge dataset of a beetles counting experiment with the following exemplary structure:
species_name1 <- c("A", "A", "A", "A", "B") # two factors for name1
species_name2 <- c("a", "a", "b", "b", "c") # three factors for name2
date <- c("2021-06-02", "2021-08-20", "2021-06-15", "2021-08-20", "2021-08-20") # three date factors
number <- c("30", "30", "11", "15", "40") # number of encountered beetles for the "date"
df <- data.frame(species_name1, species_name2, date, number) # create dataframe
df$species_full_name <- gsub(" ", " ", paste(df$species_name1, df$species_name2)) # new column with merged data of the first two columns
df$date <- as.Date(df$date, format ="%Y-%m-%d")
df$number <- as.numeric(df$number)
df$species_name1 <- as.factor(df$species_name1)
df$species_name2 <- as.factor(df$species_name2)
df$species_full_name <- as.factor(df$species_full_name)
str(df)
Overall there are three date factors (2021-06-02, 2021-06-15, 2021-08-20), but not for every "species_full_name". I need to create a dataframe which includes every single of the three dates for the factors of the "species_full_name" column. For "species_full_name"-factors with not existing "date" in the originally dataframe dates R should write a '0' to the column "numbers".
I found a code which is nearly a solution for my target dataframe. The problem is that the other columns ("species_name1" and..."_name2") will disappear:
as.data.frame(xtabs(number ~ species_full_name+date, df)) # create every factor "date" for every factor "species_full_name" and give counting data in column "Freq"
I need a dataframe which is similar to this output, but with every column from the original dataframe "df". It's important to assume the values for the columns “species_name1” and “species_name2” too.
Thanks for your help!
You can use complete()
from tidyr
complete(df, species_full_name,date) %>%
mutate(number=if_else(is.na(number),0,number))
Output:
species_full_name date species_name1 species_name2 number
<fct> <date> <fct> <fct> <dbl>
1 A a 2021-06-02 A a 30
2 A a 2021-06-15 NA NA 0
3 A a 2021-08-20 A a 30
4 A b 2021-06-02 NA NA 0
5 A b 2021-06-15 A b 11
6 A b 2021-08-20 A b 15
7 B c 2021-06-02 NA NA 0
8 B c 2021-06-15 NA NA 0
9 B c 2021-08-20 B c 40
However a data.table approach will be faster. You can use data.table
and CJ()
as follows:
# load library
library(data.table)
# set df as data.table
setDT(df)
# get unique values of species_full_name and date
species_full_name = unique(df$species_full_name)
date = unique(df$date)
# merge (and update number to 0 if NA, and the name1 and name2 columns)
merge(CJ(date,species_full_name),df,by=c('date','species_full_name'),all.x = T) %>%
.[, number:=fifelse(is.na(number),0,as.double(number))] %>%
.[, c("species_name1","species_name2"):=tstrsplit(species_full_name, " ")] %>%
.[]
Output:
date species_full_name species_name1 species_name2 number
<Date> <fctr> <char> <char> <num>
1: 2021-06-02 A a A a 30
2: 2021-06-02 A b A b 0
3: 2021-06-02 B c B c 0
4: 2021-06-15 A a A a 0
5: 2021-06-15 A b A b 11
6: 2021-06-15 B c B c 0
7: 2021-08-20 A a A a 30
8: 2021-08-20 A b A b 15
9: 2021-08-20 B c B c 40
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.