How to convert a *.txt file (copy/pasted variables) into a tabular format - regex

Question

I have a bunch of variables (roughly 80) which I copy+paste into my editor (get those variables from a different *.txt file). After this, it looks a bit messy like
ka15 1-2 tre15 3-4 hsha15 5
juso15 6
kl15 7-9 kkjs15 10
but I'd like to have it structured to get a better idea of what's going on inside the code. I also have to strip away the 15 from each variable. Ideally I would get something like
ka 1-2 tre 3-4 hsha 5
juso 6 kl 7-9 kkjs 10
Is there a clever way to achieve this? I am using SAS Enterprise Guide Editor and VSCode but couldn't find a way. For Example, when I find and replace the 15 I would wish I could replace it with a tab, but couldn't find that option in neither editors. Any ideas to get this automated or at least not do everything by hand?

Answer 1

I found a hacky solution to your problem, if anyone finds a better solution, I'll delete mine, but here it goes ¯\_(ツ)_/¯:
1) Copy all content of file(for example I copied yours twice):
ka15 1-2 tre15 3-4 hsha15 5
juso15 6
kl15 7-9 kkjs15 10
ka15 1-2 tre15 3-4 hsha15 5
juso15 6
kl15 7-9 kkjs15 10
2) Ctrl+H and replace all 15 with nothing (leave empty) using Ctrl+Alt+Enter.
3) Ctrl+F and turn Regular expressions in search box. Now type \s to select whitespace and it should select one whitespace after every word. Now select all occurrences with Alt+Enter and press Backspace followed by Enter. This will delete spaces between the words and place one word on one line of code like so:
ka
1-2
tre
3-4
hsha
...
Press Escape to remove multiple cursors.
4) Press Ctrl+F again and in search box type $ sign. This wil select end of every line. Again, press Alt+Enter to select all occurrences a press Space 5-8 times. Notice however that cursors are not properlly aligned. Press Escape to remove multiple cursors.
5) Place cursor a few spaces from a first word. Then, hold Ctrl+Alt+↓ to add multiple cursors below first one. Then, press Shift+End to select all the whitespace to the end of every line and press Delete to delete it. Press Delete again to align all words in one line seperated by n spaces.
6) Unfortunately, I couldn't find regex for the last part. Cursor should be placed after every 6th variable, but I solved it with by placing cursor next to every 7th word and pressing Enter.
I usually don't type too much like this, but I liked the problem you had. It was more puzzle than a problem to me.

Answer 2

I've come up with 3 regex's that will do what you want. In order to run them all sequentially you will need the regreplace extension or similar.
This goes in your settings:
"regreplace.on-save": false,
"regreplace.commands": [
{
"name": "Transform Data to Table Format, step 1",
"regexp": "([a-zA-Z]+|[\\d-]+)(15)?(\\s[\r\n]?)*",
"replace": "$1 \n",
"priority": 1,
},
{
"name": "Transform Data to Table Format, step 2",
"regexp":
"(([\\S-] {6})(.*))|(([\\S-]{2} {5})(.*))|(([\\S-]{3} {4})(.*))|(([\\S-]{4} {3})(.*))",
"replace": "$2$5$8$11",
"priority": 2,
},
{
"name": "Transform Data to Table Format, step 3",
"regexp":
"((.*)\n)((.*)\n)((.*)\n)((.*)\n)((.*)\n)((.*?)(\\s*\\n))",
"replace": "$2$4$6$8$10$12\n",
"priority": 3,
}
],
It creates a rule for each of the three regex steps. All three rules can be run sequentially by running the regreplace.regreplace command. Here is a demo:
The regex's are designed to look good with data items up to 4 characters long but could be easily modified for longer items.
In step 1, increase the number of spaces before the \n in the replace rule to something like 16 or so.
In step 2, you will have to sense the pattern of the regex groups like (([\\S-]{4} {3})(.*) to modify them. A 13 character long variable might require something like (([\\S-]{13} {3})(.*) as the last group and ([\\S-] {15})(.*))as the first in the sequence, etc. modifying all the other groups in order. Let me know if you need help with that.
Step 3 needs no modification unless you want to change how many data items appear on each line - right now there are 3 variables with their data on each line hence 6 groups in that regex.
It does not matter how many data-value pairs are in any row prior to running the command.
[Two items of caution: There should not be any empty lines before the start of the data, although if necessary you could a regex as the first rule to remove empty lines. Empty lines within the data or at the end are not a problem.
Secondly, the extension cannot be run on selected text only so you will have to place your data at the top of an empty file to convert it and then copy it elsewhere if you wish.]
There is also the replace rules extension which works like regreplace but will according to the docs run on a selection only but it didn't work for me here for some unknown reason. It does have a nicer interface though - all regex's could go into a single rule which could then be independently run.