How to format given string using regex? - regex

Question

So I have defined variables in such a way in my file:
public static final String hello_world = "hello world"
public static final String awesome_world = "awesome world"
public static final String bye_world= "bye world"
I have many declarations like that.
Is it possible to format them as(All '=' in a line):
public static final String hello_world = "hello world"
public static final String awesome_world = "awesome world"
public static final String bye_world = "bye world"
I can't even think of a way to do it. Any kind of help is appreciated.
P.S If it matters, I use sublime text 2.

Answer 1

If it is a one-time task you might try the following:
Import the text file into, e.g., Excel using the 'text in columns' functionality (separation character: space) so that column A contains "public" in each row, column B "static", ..., column E the variable names, column F the "=" signs, and column G the variable values (strings).
Then put the following formula into cell H1 (and copy it down to the other rows):
="public static final String "&E1&REPT(" ";50-LEN(E1))&" = "&""""&G1&""""
Afterwards, column H contains the following outputs:
public static final String hello_world = "hello world"
public static final String awesome_world = "awesome world"
public static final String bye_world = "bye world"
Please note that the Excel functions REPT and LEN are named differently if your Excel language is not English.

Answer 2

If you're careful with your original layout (so that = signs are separated from the variable name, for example, unlike the third line of data in the example), then this will do the job:
awk '{ if (length($5) > max) max = length($5);
name[NR] = $5; value[NR] = $0; sub(/^[^"]*"/, "\"", value[NR]); }
END { format = sprintf("public static final String %%-%ds = %%s\n", max);
for (i = 1; i <= NR; i++) printf(format, name[i], value[i]); }'
It assumes you are dealing with 'public static final String' throughout (but doesn't verify that). It keeps track of the length of the longest name it reads (line 1), and also the variable name and the material from the open double quote to the end of line (line 2). At the end, it generates a format string which will print the variable names left justified in a field as long as the longest (line 3). It then applies that to the saved data (line 4), generating:
public static final String hello_world = "hello world"
public static final String awesome_world = "awesome world"
public static final String bye_world = "bye world"
To make it bomb-proof (e.g. the original data), you have to work a bit harder, though it shouldn't be insuperable. The simplest fix for the sloppy original format would be to pre-filter the data with:
sed 's/=/ = /'
Extra spaces around properly spaced input won't affect the output, and the missing spaces in the 3 sample line of data are fixed. It would be fiddly to do that inside awk because you'd want it to resplit the line after editing it. You could do something very similar in Perl.
Given that the volumes of data to be processed are unlikely to be in the megabyte range, let alone larger, the two command-command solution is perfectly reasonable; you're unlikely to be able to measure the cost of the sed process.

Answer 3

There is no single regex that can solve your problem. Your only option would be to run a series of regexes, one to handle each line length:
s/^(.{40})=/\1 =/
s/^(.{39})=/\1 =/
s/^(.{38})=/\1 =/
And even then, that's probably not what you want and it's probably much, much easier to do it by hand.
The problem is that the only way a regex substitution can insert different strings at different times is if what it's inserting is a backref, and there's no backref to give you your 5 - N space characters. Your other option would be to try to capture a variable number of characters, but in this case there's no way to make that do it for you either.
Regexes were not made to do things like that (they don't support arithmetic), but some text editors are, so just find a fancy text editor or do it by hand.

Answer 4

Since you're using Sublime Text 2, there's a much easier way to do that.
There's a great package for Sublime Text 2 which will do exactly what you want:
Sublime Alignment
Dead-simple alignment of multi-line selections and
multiple selections for Sublime Text 2.
Features:
Align multiple selections to the same column by inserting spaces (or
tabs)
Align all lines in a multi-line selection to the same indent
level
Align the first = on each line of a multi-line selection to the
same column
Before:
After: