REGEX: LineStrings (Geolocation) as named capture groups and also match line prefix

Question

I have an input of lines of text, of which some contain relevant GeoData of Objects. I want to identify the relevant lines by matching a given prefix (that identifies the following coordinates as belonging to the desired Geo-Object).

It may look similar to the following:

/line with irrelevant prefix
/line with irrelevant prefix, potentially also containing coordinates
[relevant prefix][bunch of characters][1 to X coordinates in the form "lat: X.XXXXXX, lon:X.XXXXXX"][bunch of other characters][other potentially relevant information]
/line with irrelevant prefix
/line with irrelevant prefix

The actual data I want to extract is in the form of a String of coordinates (LineString), in order to generate an object representing the LineString for further use in my C# code. Additional attributes (such as a name or an ID for example) might be relevant, too. However, I also want to disqualify lines that may contain coordinates, but do not include my relevant prefix.

From what I understand, I can use named capturing groups in Regexes to get substrings as "variables" from the relevant lines, like this (please do not mind the imprecise format for the coordinate):

(lat=(?<lat>\d{0,2}[.,]\d+)), (lon=(?<lon>\d{0,2}[.,]\d+))

However, as far as I can see, I cannot have my expressions match an arbitrary number of coordinates in the line (since I do not know the length of the LineString object) and at the same time have the expression match the prefix pattern.

Is there a solution to have the expression match the prefix, have named capturing groups for the arbitrarily many pairs of coordinates, and also have additional named capturing groups for potentially relevant additional variables?

Suppose I have the following line:

PREFIXafkzh(lat=34.42344, lon=23.6346jsdfkh,lat=2.4234, lon=12.124)

I have tried the following regex:

(lat=(?<lat>\d{0,2}[.,]\d+)), (lon=(?<lon>\d{0,2}[.,]\d+))

This matches all coordinates, but not the prefix

(PREFIX).\*(lat=(?<lat>\d{0,2}[.,]\d+)), (lon=(?<lon>\d{0,2}[.,]\d+))
(PREFIX).\*((lat=(?<lat>\d{0,2}[.,]\d+)), (lon=(?<lon>\d{0,2}[.,]\d+)))+

Both of these will only match the last coordinate in the line.

the second coord prefix is inside the (), the first one isn't in this case — Schwarzie2478, Feb 20 '23 at 13:45
I overlooked that StackOverflow required some more escapes for the regexes, they should make more sense now. However, I am not sure I understand your question. The prefix is a separate capture group, which doesn't necessarily need to be named, though. — Jannik Michel, Feb 20 '23 at 13:49
I'm confused about the string jsdfkh, is it part of the first or second coordinate? if second, then i would expect afksh to be part of the first coordinate, but it's not part of the () — Schwarzie2478, Feb 20 '23 at 14:06

Patrick Janser · Accepted Answer · 2023-02-20T14:59:23.070

It's not very easy to know what kind of prefixes you have and if there's one for the line and another prefix just for a group of coordinates. You may have to add a few more examples of input data.

I understood that you have a prefix for the line followed by a ( then some coordinates that could also be prefixed by a word and a comma.

Assuming that, I came out with this .NET regex:

@"(?<line_prefix>\w+)\(          # Line word prefix followed by (.
(?:                              # Non-capturing group to match multiple coords.
  (?<coord_prefix>[\w,]*)        # Optional coord prefix (may have to adapt).
  (?:lat=(?<lat>\d{0,2}[.,]\d+)) # Latitude value.
  ,\s*                           # Comma and optional spaces.
  (?:lon=(?<lon>\d{0,2}[.,]\d+)) # Longitude value.
)+"gx

You can test it here: https://regex101.com/r/vxjeww/3

You may have to adapt the line prefix and coordinate prefix pattern as we don't know clearly what it can contain.

As you see, you should then get multiple variables as the non-capturing group can be matched 1 or several times.

Yes, thank you! The use of non-capturing groups seems to have been the missing link! — Jannik Michel, Feb 20 '23 at 15:11

REGEX: LineStrings (Geolocation) as named capture groups and also match line prefix

1 Answers1