Varying results when re-running same query

I ran a simple query a total of four times immediately after the previous one had finished (scraper output deleted for clarity):

```
$ echo "dentists in nuneaton" > ~/Downloads/test.txt
$ google_maps_scraper -input ~/Downloads/test.txt -results ~/Downloads/test-1.json -json
$ google_maps_scraper -input ~/Downloads/test.txt -results ~/Downloads/test-2.json -json
$ google_maps_scraper -input ~/Downloads/test.txt -results ~/Downloads/test-3.json -json
$ google_maps_scraper -input ~/Downloads/test.txt -results ~/Downloads/test-4.json -json
```

From each JSON output file I then grepped/sorted the `place_id` and compared the results. To my surprise, of a total of 39 scraped places, only 21 (= 54%) were present in all runs – see attached CSV.

The total run time of all four jobs was just over two minutes, so additions to/deletions from Google Maps can safely be assumed to not have happened.

Is this an inherent limitation of Google’s DB response or a problem with the scraper?

[comparison.csv](https://github.com/user-attachments/files/28957338/comparison.csv)