# Read then write to same file for code (text file) formatting
**URL:** https://discourse.julialang.org/t/read-then-write-to-same-file-for-code-text-file-formatting/115095
**Category:** General Usage
**Tags:** question, io
**Created:** [June 3, 2024, 3:57am UTC](https://discourse.julialang.org/t/read-then-write-to-same-file-for-code-text-file-formatting/115095 "2024-06-03T03:57:14Z")
**Posts on this page:** 6
**Page:** 1
### Author:  [@paalon](https://discourse.julialang.org/u/paalon)
#### Post date: [June 3, 2024, 3:57am UTC](https://discourse.julialang.org/t/read-then-write-to-same-file-for-code-text-file-formatting/115095/1 "2024-06-03T03:57:14Z")
I want to sort a TOML file content (I know TOML does not ensure its ordering but let’s put it aside). So, I write the following script try to read a TOML file then sort by keys and then replace it by sorted TOML file. But `"r+"` and `"w+"` options don’t work well.
```julia
using TOML
using OrderedCollections
function sortfile(filename)
open(filename, "r+") do io
unordered = TOML.parse(io)
ordered = OrderedDict(sort(collect(unordered), by=x->x[1]))
TOML.print(io, ordered)
end
end
# sort by keys
sortfile("sample.toml")
```
Generally, what kind of processing is recommended for such work?
---
### Author:  [@ufechner7](https://discourse.julialang.org/u/ufechner7)
#### Post date: [June 3, 2024, 4:54am UTC](https://discourse.julialang.org/t/read-then-write-to-same-file-for-code-text-file-formatting/115095/2 "2024-06-03T04:54:15Z")
First read the file and close it, then write it.
I use:
```julia
function readfile(filename)
open(filename) do file
readlines(file)
end
end
function writefile(lines, filename)
open(filename, "w") do file
for line in lines
write(file, line, '\n')
end
end
end
```
---
### Author:  [@paalon](https://discourse.julialang.org/u/paalon)
#### Post date: [June 3, 2024, 6:56am UTC](https://discourse.julialang.org/t/read-then-write-to-same-file-for-code-text-file-formatting/115095/3 "2024-06-03T06:56:21Z")
DaveMacMahon taught me I need `seekstart(io)` and `truncate(io, position(io))` before `TOML.print`.
```julia
using TOML
using OrderedCollections
function sortfile(filename)
open(filename, "r+") do io
unordered = TOML.parse(io)
OrderedDict(unordered)
ordered = OrderedDict(sort!(collect(unordered), by=first))
seekstart(io)
truncate(io, position(io))
TOML.print(io, ordered)
end
end
# sort by keys
sortfile("sample.toml")
```
---
### Author:  [@paalon](https://discourse.julialang.org/u/paalon)
#### Post date: [June 3, 2024, 6:58am UTC](https://discourse.julialang.org/t/read-then-write-to-same-file-for-code-text-file-formatting/115095/4 "2024-06-03T06:58:56Z")
Thanks, but I want to do it only opening file once.
---
### Author:  [@GunnarFarneback](https://discourse.julialang.org/u/GunnarFarneback)
#### Post date: [June 3, 2024, 7:20am UTC](https://discourse.julialang.org/t/read-then-write-to-same-file-for-code-text-file-formatting/115095/5 "2024-06-03T07:20:25Z")
Is opening files very slow for you or do you just not want to call `open` twice? If it’s the latter you can hide the opening for reading with
```julia
function sortfile(filename)
unordered = TOML.parsefile(filename)
ordered = OrderedDict(sort!(collect(unordered), by=first))
open(filename, "w") do io
TOML.print(io, ordered)
end
end
```
---
### Author:  [@barucden](https://discourse.julialang.org/u/barucden)
#### Post date: [June 3, 2024, 8:42am UTC](https://discourse.julialang.org/t/read-then-write-to-same-file-for-code-text-file-formatting/115095/6 "2024-06-03T08:42:23Z")
Your question was what kind of processing is recommended. I would say that it is recommended that, while reading from file `A`, you write into a temporary file `B`. Once you are done writing (and everything succeeded) you rename `B` to `A`.