Regex Return First Match - json

Question

I have a weather file where I would like to extract the first value for "air_temp" recorded in a JSON file. The format this HTTP retriever uses is regex (I know it is not the best method).
I've shortened the JSON file to 2 data entries for simplicity - there are usually 100.
{
"observations": {
"notice": [
{
"copyright": "Copyright Commonwealth of Australia 2017, Bureau of Meteorology. For more information see: http://www.bom.gov.au/other/copyright.shtml http://www.bom.gov.au/other/disclaimer.shtml",
"copyright_url": "http://www.bom.gov.au/other/copyright.shtml",
"disclaimer_url": "http://www.bom.gov.au/other/disclaimer.shtml",
"feedback_url": "http://www.bom.gov.au/other/feedback"
}
],
"header": [
{
"refresh_message": "Issued at 12:11 pm EST Tuesday 11 July 2017",
"ID": "IDN60901",
"main_ID": "IDN60902",
"name": "Canberra",
"state_time_zone": "NSW",
"time_zone": "EST",
"product_name": "Capital City Observations",
"state": "Aust Capital Territory"
}
],
"data": [
{
"sort_order": 0,
"wmo": 94926,
"name": "Canberra",
"history_product": "IDN60903",
"local_date_time": "11/12:00pm",
"local_date_time_full": "20170711120000",
"aifstime_utc": "20170711020000",
"lat": -35.3,
"lon": 149.2,
"apparent_t": 5.7,
"cloud": "Mostly clear",
"cloud_base_m": 1050,
"cloud_oktas": 1,
"cloud_type_id": 8,
"cloud_type": "Cumulus",
"delta_t": 3.6,
"gust_kmh": 11,
"gust_kt": 6,
"air_temp": 9.0,
"dewpt": 0.2,
"press": 1032.7,
"press_qnh": 1031.3,
"press_msl": 1032.7,
"press_tend": "-",
"rain_trace": "0.0",
"rel_hum": 54,
"sea_state": "-",
"swell_dir_worded": "-",
"swell_height": null,
"swell_period": null,
"vis_km": "10",
"weather": "-",
"wind_dir": "WNW",
"wind_spd_kmh": 7,
"wind_spd_kt": 4
},
{
"sort_order": 1,
"wmo": 94926,
"name": "Canberra",
"history_product": "IDN60903",
"local_date_time": "11/11:30am",
"local_date_time_full": "20170711113000",
"aifstime_utc": "20170711013000",
"lat": -35.3,
"lon": 149.2,
"apparent_t": 4.6,
"cloud": "Mostly clear",
"cloud_base_m": 900,
"cloud_oktas": 1,
"cloud_type_id": 8,
"cloud_type": "Cumulus",
"delta_t": 2.9,
"gust_kmh": 9,
"gust_kt": 5,
"air_temp": 7.3,
"dewpt": 0.1,
"press": 1033.1,
"press_qnh": 1031.7,
"press_msl": 1033.1,
"press_tend": "-",
"rain_trace": "0.0",
"rel_hum": 60,
"sea_state": "-",
"swell_dir_worded": "-",
"swell_height": null,
"swell_period": null,
"vis_km": "10",
"weather": "-",
"wind_dir": "NW",
"wind_spd_kmh": 4,
"wind_spd_kt": 2
}
]
}
}
The regex expression I am currently using is: .*air_temp": (\d+).* but this is returning 9 and 7.3 (entries 1 and 2). Could someone suggest a way to only return the first value?
I have tried using lazy quantifier group, but have had no luck.

Answer 1

This regex will help you. But I think you should capture and extract the first match with features of the programming language you are using.
.*air_temp": (\d{1,3}\.\d{0,3})[\s\S]*?},
To understand the regex better: take a look at this.
Update
The above solution works if you have only two data entries. For more than two entries, we should have used this one:
header[\s\S]*?"air_temp": (\d{1,3}\.\d{0,3})
Here we match the word header first and then match anything in a non-greedy way. After that, we match our expected pattern. thus we get the first match. Play with it here in regex101.
To capture the negative numbers, we need to check if there is any - character exists or not. We do this by ? which means 'The question mark indicates zero or one occurrence of the preceding element'.
So the regex becomes,
header[\s\S]*?"air_temp": (-?\d{1,3}\.\d{0,3}) Demo
But the use of \K without the global flag ( in another answer given by mickmackusa ) is more efficient. To detect negative numbers, the modified version of that regex is
air_temp": \K-?\d{1,2}\.\d{1,2} demo.
Here {1,2} means 1~2 occurance/s of the previous character. We use this as {min_occurance,max_occurance}

Answer 2

I do not know which language you are using, but it seems like a difference between the global flag and not using the global flag.
If the global flag is not set, only the first result will be returned. If the global flag is set on your regex, it will iterate through returning all possible results. You can test it easily using Regex101, https://regex101.com/r/x1bwg2/1
The lazy/greediness should not have any impact in regards to using/not using the global flag

Answer 3

If \K is allowed in your coding language, use this: Demo
/air_temp": \K[\d.]+/ (117steps) this will be highly efficient in searching your very large JSON text.
If no \K is allowed, you can use a capture group: (Demo)
/air_temp": ([\d.]+)/ this will still move with decent speed through your JSON text
Notice that there is no global flag at the end of the pattern, so after one match, the regex engine stops searching.
Update:
For "less literal" matches (but it shouldn't matter if your source is reliable), you could use:
Extended character class to include -:
/air_temp": \K[\d.-]+/ #still 117 steps
or change to negated character class and match everything that isn't a , (because the value always terminates with a comma):
/air_temp": \K[^,]+/ #still 117 steps
For a very strict match (if you are looking for a pattern that means you have ZERO confidence in the input data)...
It appears that your data doesn't go beyond one decimal place, temps between 0 and 1 prepend a 0 before the decimal, and I don't think you need to worry with temps in the hundreds (right?), so you could use:
/air_temp": \K-?[1-9]?\d(?:\.\d)? #200steps
Explanation:
Optional negative sign
Optional tens digit
Required ones digit
Optional decimal which must be followed by a digit
Accuracy Test Demo
Real Data Demo