How can I remove the prefix (index indicator) [1] in knitr output?

Question

Standard R output looks like this

> 3
[1] 3

To remove the prefix 1 you can use

> cat(3)
3

Is there a way to remove this globally? Or do you have to wrap cat() around everything?

Further to that, I'm using this within knitr, so if there isn't an R global setting, there may be a knitr wide setting, I did look, but couldn't see one.

Edit: It was asked why one would want this, something like if you wanted to structure a report like the below. The [1] is just not needed and means nothing to none R users (aka, the audience).

Additional Info: In knitr you can use \Sexpr{} or r ... to evaluate something in line, and in that scenario it doesn't print the [1]. For example if you had:

There are `r sum(mtcars$cyl==6)` cars with 6 cylinders.

You would get:

There are 7 cars with 6 cylinders.

As your output, not:

There are [1] 7 cars with 6 cylinders.

Answer 1

Nothing is impossible. Take a look at what can be done with knitr hooks.
Have fun!
_{^{Rmarkdown script gist}}

# A Prefix nulling hook.

# Make sure to keep the default for normal processing.
default_output_hook <- knitr::knit_hooks$get("output")

# Output hooks handle normal R console output.
knitr::knit_hooks$set( output = function(x, options) {

  comment <- knitr::opts_current$get("comment")
  if( is.na(comment) ) comment <- ""
  can_null <- grepl( paste0( comment, "\\s*\\[\\d?\\]" ),
                     x, perl = TRUE)
  do_null <- isTRUE( knitr::opts_current$get("null_prefix") )
  if( can_null && do_null ) {
    # By default R print output aligns at the right brace.
    align_index <- regexpr( "\\]", x )[1] - 1
    # Two cases: start or newline
    re <- paste0( "^.{", align_index, "}\\]")
    rep <- comment
    x <- gsub( re, rep,  x )
    re <- paste0( "\\\n.{", align_index, "}\\]")
    rep <- paste0( "\n", comment )
    x <- gsub( re, rep,  x )
  }

  default_output_hook( x, options )

})

knitr::opts_template$set("kill_prefix"=list(comment=NA, null_prefix=TRUE))

Normal:

```{r}
print( 1:50 )
```

##  [1]  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
## [24] 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
## [47] 47 48 49 50

Null prefix

```{r, null_prefix=TRUE}
print( 1:50 )
```

##  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
## 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
## 47 48 49 50

Set as default (and remove comment string):

```{r}
knitr::opts_chunk$set(opts.label="kill_prefix")
```

```{r}
print( 1:50 )
```

  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
 47 48 49 50

Ensure we aren't killing strings with [:digit:] patterns.

```{r}
print( paste0( paste0("[", 1:50), "]" ),quote = FALSE)
```

 [1]  [2]  [3]  [4]  [5]  [6]  [7]  [8]  [9]  [10] [11] [12] [13] [14]
 [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]
 [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42]
 [43] [44] [45] [46] [47] [48] [49] [50]

Answer 2

Can't be done. But I would hesitate to either do it, or just wrap everything in a cat. Why?

Firstly, R users expect to see the [1] label, and its helpful when the vector is longer than one line, or if its a matrix with rows and column labels. Why do you want to do this?

Secondly, now you know why you want to do this, write a function named after the reason you want to do this, that does this, and wrap your outputs in that function. Because I reckon if you want fine control over the output you are pretty soon going to go "Hmmm cat isn't quite cutting it for me, I want to change this and that and this and that" and that's easier if you have wrapped your output in a function you can change. At first it might just be cleanprint=function(...){cat(...)} but then its easier to move on from there by redefining cleanprint than tweaking cat. You might want to use sprintf, or set the number of decimal digits or something.