Remove whitespace after +44 and quotation marks around lat/long in JSON string with regex -

in example snippet below have json needs edited (over 1400 entries). need achieve 2 things:

1. in example line: "phone": "+44 2079693900", need remove whitespace between +44 , 2079693900 records. resulting in: "+442079693900"

2. for latitude , longitude need rid of double quotes around numbers, api using accepts these values floats. example: "latitude": "51.51736", needs be: "latitude": 51.51736

i familiar ruby, , have done parsing of json in past, thought regex best tool use kind of basic data cleaning task. have referred regex101.com , regular-expressions.info i'm pretty stuck @ point. in advance!

[   {     "id": "101756",     "name": "1 lombard street     "email": "reception@1lombardstreet.com",     "website": "http://www.1lombardstreet.com",     "location": {       "latitude": "51.5129",       "longitude": "-0.089",       "address": {         "line1": "1 lombard street",         "line2": "",         "line3": "",         "postcode": "ec3v 9aa",         "city": "london",         "country": "uk"       }     }   },   {     "id": "105371",     "name": "108 brasserie",     "phone": "+44 2079693900",     "email": "enquiries@108marylebonelane.com",     "website": "http://www.108brasserie.com",     "location": {       "latitude": "51.51795",       "longitude": "-0.15079",       "address": {         "line1": "108 marylebone lane",         "line2": "",         "line3": "",         "postcode": "w1u 2qe",         "city": "london",         "country": "uk"       }     }   },   {     "id": "108701",     "name": "1901 restaurant",     "phone": "+44 2076187000",     "email": "london.restres@andaz.com",     "website": "http://www.andazdining.com",     "location": {       "latitude": "51.51736",       "longitude": "-0.08123",       "address": {         "line1": "andaz hotel",         "line2": "40 liverpool street",         "line3": "",         "postcode": "ec2m 7qn",         "city": "london",         "country": "uk"       }     }   },   {     "id": "102190",     "name": "2 bridge place",     "phone": "+44 2078028555",     "email": "fb@dtlondonvictoria.com",     "website": "http://crimsonhotels.comdoubletreelondonvictoriadiningpre-theatre-dining",     "location": {       "latitude": "51.49396",       "longitude": "-0.14343",       "address": {         "line1": "2 bridge place",         "line2": "victoria",         "line3": "",         "postcode": "sw1v 1qa",         "city": "london",         "country": "uk"       }     }   },   {     "id": "102063",     "name": "2 veneti",     "phone": "+44 2076370789",     "email": "2veneti@btconnect.com",     "website": "http://www.2veneti.com",     "location": {       "latitude": "51.5168",       "longitude": "-0.14673",       "address": {         "line1": "10 wigmore street",         "line2": "",         "line3": "",         "postcode": "w1u 2rd",         "city": "london",         "country": "uk"       }     }   }, 

you can use following regex:

("phone":\s*"\+44)\s+|("(?:latitude|longitude)":\s*)"([^"]+)" 

with following replacement:

$1$2$3 

the idea capture want , not capture not, , use backreferences restore substrings want keep.

regex explanation:

the pattern contains 2 alternatives joined | alternation operator:

1. ("phone":\s*"\+44)\s+:
   • ("phone":\s*"\+44) - 1st capturing group matching literal "phone": + optional whitespace, +44 literally
   • \s+ - 1 or more whitespaces we'll remove
2. ("(?:latitude|longitude)":\s*)"([^"]+)":
   • ("(?:latitude|longitude)":\s*) - second capturing group matching "latitude": or "longitude": , 0 or more whitespace characters
   • " - literal " we'll drop
   • ([^"]+) - third capturing group matching 1 or more characters other " (we'll keep that)
   • " - again, literal " we'll drop.

see demo