flow_control2.Rmd

---
title: "More Loops and Conditionals"
output:
  html_document:
    df_print: paged
---

### Using multiple flow controls

Loops can get really complicated really fast when depending on the complexity of your question. For example here is how to find the biggest thing in a simple data frame (in a really inefficient way). 

```{r}

loop.df <- data.frame(lets = LETTERS[1:5], 
                      nums = c(10, 3, 36, 12 , 9))

x <- NULL
y <- NULL
for(i in 1:nrow(loop.df)){
  
  if(is.null(x)){
    x <- loop.df[i,1]
    y <- loop.df[i,2]
  }else{
    if(y < loop.df[i,2]){
      x <- loop.df[i,1]
      y <- loop.df[i,2]
    }
  }  
}

x
```

Now for only odd numbers.

```{r}

loop.df <- data.frame(lets = LETTERS[1:5], 
                      nums = c(10, 3, 36, 12 , 39))

x <- NULL
y <- NULL

for(i in 1:nrow(loop.df)){
      
    if(loop.df[i,2] %% 2 != 0 ){
      if(is.null(x)){
        x <- loop.df[i,1]
        y <- loop.df[i,2]
      }else{ 
        if(y < loop.df[i,2]){
          x <- loop.df[i,1]
          y <- loop.df[i,2]
        } 
    } 
  }
}

x
```
Now for only odd numbers less than 30.

```{r}

loop.df <- data.frame(lets = LETTERS[1:5], 
                      nums = c(10, 3, 36, 12 , 39))

x <- NULL
y <- NULL

for(i in 1:nrow(loop.df)){
    if(loop.df[i,2] > 30 ){
      print("too big")
      break
    }  
    if(loop.df[i,2] %% 2 != 0 ){
      if(is.null(x)){
        x <- loop.df[i,1]
        y <- loop.df[i,2]
      }else{ 
        if(y < loop.df[i,2]){
          x <- loop.df[i,1]
          y <- loop.df[i,2]
        } 
    } 
  }
}

x
```

Break can restrict the range and scope of loops

***


Just so you know a more efficient way of doing this for the first example
```{r}

loop.df <- data.frame(lets = LETTERS[1:5], 
                      nums = c(10, 3, 36, 12 , 9))

loop.df <- loop.df[order(loop.df$nums,decreasing = T),] # now reorder

loop.df[1,1]

```

For the second example ... 

```{r}

loop.df <- data.frame(lets = LETTERS[1:5], 
                      nums = c(10, 2, 36, 7 , 9))

loop.df <- loop.df[which(loop.df$nums %% 2 != 0),]

loop.df

loop.df <- loop.df[order(loop.df$nums,decreasing = T),]

loop.df

loop.df[1,1]

```
This introduces a couple new pieces of syntax and a very important concept: reduce your problem into manageable sub-problems for which a efficient solution exists. Here we use the functions `which` and `order`. Both we have previously used in the context of manipulating data frames. This illustrates the point that sometimes a loop is not the best solution to your problem, particularly if you have a large amount of data.   



#### Exercise 

Select a blast/diamond report you have previously generated that has more than 100 hits. Get the first 100 hits in the report and save them to a new file. Then import this new file into R and ***using loops and conditionals*** find the overall best hit (regardless of query or sequence) in the file based off 1) percent ID, 2) bit score, 3) evalue (three separate loops). Next design a loop that will 1) first check to ensure that the total length the alignment is at least 95% of the shorter sequence, then 2) return the best hit (percent ID) that passes this test.