Logic parsing in R - r

Question

Building off of this question Pass a data.frame with column names and fields as filter
Let's say we have the following data set:
filt = data.table(X1 = c("Gender","Male"),
X2 = c('jobFamilyGroup','Finance'),
X3 = c('jobFamilyGroup','Software Dev')
df = data.table(Gender = c('Male','F','Male','Male','F'),
EmployeeStatus = c('Active','na','Active','Active','na'),
jobFamilyGroup = c('Finance','Software Dev','HR','Finance','Software Dev'))
and I want to use filt as a filter for df. filt is done by grabbing an input from Shiny and transforming it a bit to get me that data.table above. My goal is to filter df so we have: All rows that are MALE AND (Software Dev OR Finance).
Currently, I'm hardcoding it to always be an AND but that isn't ideal for situations like this. My thought would be to have multiple if conditions to catch things like this, but I feel like there could be an easier approach for building this logic in.
___________UPDATE______________
Once I have a table like filt I can pass code like:
if(!is.null(primary))
{
if(ncol(primary)==1){
d2 = df[get(as.character(primary[1,1]))==as.character(primary[2,1])]
}
else if(length(primary)==2){
d2 = df[get(as.character(primary[1,1]))==as.character(primary[2,1]) &
get(as.character(primary[1,2]))==as.character(primary[2,2])]
}
else{
d2 = df[get(as.character(primary[1,1]))==as.character(primary[1,2]) &
get(as.character(primary[1,2]))==as.character(primary[2,2]) &
get(as.character(primary[1,3]))==as.character(primary[2,3])]
}
}
But this code doesn't account for the OR Logical needed if there are multiple inputs for one type of grouping. Meaning the current code says give me all rows where: Gender == Male & Job Family Group == 'Finance'& Job Family Group == 'Software Dev' When really it should be Gender == Male & (Job Family Group == 'Finance'| Job Family Group == 'Software Dev')
this is a minimal example meaning there are many other columns so ideally the solution has the ability to determine when a multiple input for a grouping is present.

Answer 1

Given your problem, what if you parsed it so your logic looked like:
Gender %in% c("Male") & jobFamilyGroup %in% c('Finance','Software Dev')
By lumping all filter values with the same column name together in an %in% you get your OR and you keep your AND between column names.
UPDATE
Consider the case discussed in comments below.
Your reactive inputs a data.table specifying
Gender IS Male
Country IS China OR US
EmployeeStatus IS Active
In the sample data you provided there is no country column, so I added one. I extract the columns to be filtered and the values to be filtered and split the values to be filtered by the columns. I pass this into an lapply which does the logical check for each column using an %in% rather than a == so that options within the same column are treated as an | instead of a &. Then I rbind the logical results together and apply an all to the columns and then filter df by the results.
This approach handles the & between columns and the | within columns. It supports any number of columns to be searched removing the need for your if/else logic.
library(data.table)
df = data.table(Gender = c('Male','F','Male','Male','F'),
EmployeeStatus = c('Active','na','Active','Active','na'),
jobFamilyGroup = c('Finance','Software Dev','HR','Finance','Software Dev'),
Country = c('China','China','US','US','China'))
filt = data.table(x1 = c('Gender' , 'Male'),x2 = c('Country' , 'China'),x3 = c('Country','US'), x4 = c('EmployeeStatus','Active'))
column = unlist(filt[1,])
value = unlist(filt[2,])
tofilter = split(value,column)
tokeep = apply(do.call(rbind,lapply(names(tofilter),function(x){
`[[`(df,x) %in% tofilter[[x]]
})),2,all)
df[tokeep==TRUE]
#> Gender EmployeeStatus jobFamilyGroup Country
#> 1: Male Active Finance China
#> 2: Male Active HR US
#> 3: Male Active Finance US