enter image description here
A text version of what my dataset looks like
ID Sample_Type Species
1 Sample A
1 Sample B
1 Sample C
1 Sample D
2 Sample A
2 Sample B
2 Sample C
2 Sample D
2 Sample E
2 Sample F
2 control NA
3 sample B
3 sample C
3 sample D
3 sample E
3 sample G
3 control B
4 sample A
4 sample B
4 sample C
4 sample G
4 sample J
4 sample K
5 control NA
Description
This is a basic example of my dataset. In this example I want to retain all rows for ID's 2 and 3 because they contain a sample and a control. Whereas samples 1 and 4 do not contain controls and therefore I do not need to analyze them at this time. Simularly examples like ID 5 are independent controls that are not paired with other samples, I dont want these included either.
I intend to compare the results of paired samples and controls. The issue is that the paired ID's are within a broader dataset which is very large and the paired samples/controls represent less than 10% of the overall dataset.
I am not looking to randomly sample this dataset nor keep 1 row like much of the other questions on stack. Only to clean it to where I keep every row that meets the criteria of the if statement below and remove all observations that dont meet that criteria. For example to exclude samples with no controls or independent controls that dont have samples associated with them.
data2<- data %>%
select(ID, Sample_Type, Species) %>%
filter(if (data$Sample_Type = "Control" & "Sample") {SiteID = TRUE})
enter image description here
A text version of what my dataset looks like
ID Sample_Type Species
1 Sample A
1 Sample B
1 Sample C
1 Sample D
2 Sample A
2 Sample B
2 Sample C
2 Sample D
2 Sample E
2 Sample F
2 control NA
3 sample B
3 sample C
3 sample D
3 sample E
3 sample G
3 control B
4 sample A
4 sample B
4 sample C
4 sample G
4 sample J
4 sample K
5 control NA
Description
This is a basic example of my dataset. In this example I want to retain all rows for ID's 2 and 3 because they contain a sample and a control. Whereas samples 1 and 4 do not contain controls and therefore I do not need to analyze them at this time. Simularly examples like ID 5 are independent controls that are not paired with other samples, I dont want these included either.
I intend to compare the results of paired samples and controls. The issue is that the paired ID's are within a broader dataset which is very large and the paired samples/controls represent less than 10% of the overall dataset.
I am not looking to randomly sample this dataset nor keep 1 row like much of the other questions on stack. Only to clean it to where I keep every row that meets the criteria of the if statement below and remove all observations that dont meet that criteria. For example to exclude samples with no controls or independent controls that dont have samples associated with them.
data2<- data %>%
select(ID, Sample_Type, Species) %>%
filter(if (data$Sample_Type = "Control" & "Sample") {SiteID = TRUE})
For filtering you can check if both "sample" & "control" are present in ID group, all() checks that all left side operands are present in right side of %in% and through .by we are making that check within each ID group. tolower() to handle varying capitalization in input data.
data |>
dplyr::filter(all(c("sample", "control") %in% tolower(Sample_Type)), .by = ID)
#> ID Sample_Type Species
#> 1 2 Sample A
#> 2 2 Sample B
#> 3 2 Sample C
#> 4 2 Sample D
#> 5 2 Sample E
#> 6 2 Sample F
#> 7 2 control <NA>
#> 8 3 sample B
#> 9 3 sample C
#> 10 3 sample D
#> 11 3 sample E
#> 12 3 sample G
#> 13 3 control B
Example data:
data <- read.table(header = TRUE, text = "
ID Sample_Type Species
1 Sample A
1 Sample B
1 Sample C
1 Sample D
2 Sample A
2 Sample B
2 Sample C
2 Sample D
2 Sample E
2 Sample F
2 control NA
3 sample B
3 sample C
3 sample D
3 sample E
3 sample G
3 control B
4 sample A
4 sample B
4 sample C
4 sample G
4 sample J
4 sample K
5 control NA")
data2<-data %>% group_by(ID) %>% mutate(Sample_Type = factor(Sample_Type, levels = c("Control","Sample"))) %>% filter(all(levels(Sample_Type) %in% Sample_Type)) %>% ungroup()