sumstats_by_trt.Rmd

---
title: "R Notebook"
output: html_notebook
---

```{r}
load("hw2_list.rda")
```

```{r}
calculate_stats <- function(bp_list, stat = "mean") {
    #get `stat` param from function
    stat_arg <- get(stat)
    
    #initialize return list
    sumstats <- list()
    
    # get names
    for (i in names(bp_list)) {
        df <- bp_list[[i]]
        by_trt_stats <-c()
        
        # get numeric columns
        num_vars <- names(df)[sapply(df, is.numeric)]
        
        # calculate stats for numeric variables
        for (j in num_vars) {
            col_stat <- stat_arg(df[[j]])
            stat_name <- paste(stat, j, sep = "_")
            by_trt_stats[stat_name] <- col_stat
        }
        
        # store vector in the list by group name
        sumstats[[i]] <- as.data.frame(t(by_trt_stats))
    }
    
    # return list of data frames
    return(sumstats)
}
tst_mean_def<-calculate_stats(bp_list)
```























### create a nested List from df_names as levels/parent and numeric vars and sumstats
```{r}

attributes(bp_list)
typeof(bp_list)
str(bp_list)
summary(bp_list) # bp_List is a list containing 2 data frames
is.vector(bp_list)
is.data.frame(bp_list[[1]])
names(bp_list)

get_df_names <- function(bp_list, stat = "mean"){
    stat__ <- get(stat)
    df_names <- list()
    return_list <- list()
    for(item in names(bp_list)){
        df_names <- item
        numeric_var_names <- vector()
        df_stats <- vector()
        for (j in bp_list){
            numeric_var_names <- (names(j)[sapply(j,is.numeric)])
            df_stats <- c(stat__(j$pre_bp), stat__(j$post_bp), stat__(j$diff_bp))
        }
        return_list <- append(return_list, list(df_names, numeric_var_names, df_stats))
    }
    
    return(return_list)
}
dfnamestst <- get_df_names(bp_list)
```


### Just create the 6 element list, but provide names for each stat in the list
```{r}

hobbies <- c("reading", "hiking", "coding")
hobby_string <- paste(hobbies, collapse = "_")
print(hobby_string) # Output: "reading, hiking, coding"

get_var_string <- function(bp_list, stat = "mean"){
    stat__ <- get(stat)
    df_names <- names(bp_list)
    
    return_df <- data.frame()
    
    for(i in seq_along(bp_list)){
        df_stats <- c()
        df <- bp_list[[i]]
        df_name <- df_names[i]
        
        num_vars <- names(df)[sapply(df, is.numeric)]
        
        # df_stats <- c(stat__(i$pre_bp), stat__(i$post_bp), stat__(i$diff_bp))
        
        for(col in num_vars){
            col_stat <- stat__(df[[col]])
            print(col_stat)
            var_string <- paste(df_name, col, sep= "_")
            df_stats <- append(df_stats, col_stat)
        }
        return_df <- rbind(return_df,col_stat)
    }
    return(return_df)
}
x <- get_var_string(bp_list)


create_named_list <- function(bp_list, stat = "mean"){
    stat__ <- get(stat)
    var_string <- vector()
    for(i in bp_list){
        var_string <- vector()
        var_string <- paste((names(i)[sapply(i,is.numeric)]), sep = "_")
        print(var_string)
    }
    
    
}
x2 <- create_named_list(bp_list)

# get_df_names <- function(bp_list, stat = "mean"){
#     stat__ <- get(stat)
#     df_names <- list()
#     return_list <- list()
#     for(item in names(bp_list)){
#         df_names <- item
#         numeric_var_names <- vector()
#         df_stats <- vector()
#         for (j in bp_list){
#             numeric_var_names <- (names(j)[sapply(j,is.numeric)])
#             df_stats <- c(stat__(j$pre_bp), stat__(j$post_bp), stat__(j$diff_bp))
#         }
#         return_list <- append(return_list, list(df_names, numeric_var_names, df_stats))
#     }
#     
#     return(return_list)
# }
# dfnamestst <- get_df_names(bp_list)
```
### What am I trying to do?
- iterate through a list of dataframes (bp_list), extracting:
    - names of dataframes ("treatment" and "placebo")
    - names of numeric variables
- summarize numeric columns
- label the things to make sense
- return output across trt, pbo
- return a list of dataframes containing summary stats for numeric columns
    - dataframes for treatment regimen (trt, pbo)
        - meaningfully labeled? 
        
### What is the issue?
- flow control. I really hate nesting for loops if there is a faster alternative (I don't know it)

- addressing elements of:
    - lists
    - dataframes inside of lists
```{r}
names(bp_list)
bp_list[[1]]
bp_list[1] #don't use these
bp_list$treatment #don't use these
length(bp_list)

for (item in bp_list){
    #R just knows bp_list is a list of 2 items
    # print(bp_list[item]) #why doesn't this return $treatment like it does above?
    print(item) #why does this give an invalid subscript for type 'list' error
    # print(bp_list$item) #why does this return NULL, when outside of a loop it returns a dataframe?
}
```