将OpenWeatherMap预报JSON转换为CSV适配Gnuplot的技术问询

Hey there! Nice job getting this far with json2csv—flattening the data and stripping irrelevant fields is half the battle. Let's walk through both of your requested solutions to get a clean CSV ready for Gnuplot:

Option 1: Use sed to Add Line Breaks After Plain-Text Dates #

If your current flattened output has all forecast entries crammed into one line (with plain-text dates like YYYY-MM-DD acting as entry separators), you can use sed to split them into proper CSV rows.

Example Command #

Assuming your flattened text has entries separated by dates in YYYY-MM-DD format (e.g., 2024-05-20,25,602024-05-21,26,55):

# Insert a line break before each date, then remove any leading empty line
sed -E 's/([0-9]{4}-[0-9]{2}-[0-9]{2})/\n\1/g' your_flattened_output.txt | sed '/^$/d' > formatted_forecasts.csv

Customization Tips #

• If your dates are paired with a field name (e.g., date:2024-05-20), adjust the regex to match that pattern:

  sed -E 's/(date:[0-9]{4}-[0-9]{2}-[0-9]{2})/\n\1/g' your_flattened_output.txt | sed '/^$/d' > formatted_forecasts.csv
  

• Tweak the regex to match your exact date format (e.g., add time if needed: [0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}).

Option 2: Directly Flatten JSON to CSV (No Post-Processing Needed) #

Instead of fixing a messy flattened output, use a tool like jq (more flexible than basic json2csv) or refined json2csv arguments to generate a properly formatted CSV directly from your JSON.

Using jq (Highly Recommended for Control) #

jq lets you explicitly select fields and structure the CSV perfectly for Gnuplot. Suppose your JSON looks like this:

{
  "weather_data": [
    {"date": "2024-05-20", "temp_c": 25, "humidity": 60},
    {"date": "2024-05-21", "temp_c": 26, "humidity": 55}
  ]
}

Run this command to output a CSV with headers and properly separated rows:

jq -r '["date", "temp_c", "humidity"], (.weather_data[] | [.date, .temp_c, .humidity]) | @csv' input.json > gnuplot_ready.csv

Using json2csv (Refined Arguments) #

If you prefer sticking with json2csv, use the --unwind flag to expand nested arrays and --fields to specify exactly what you want:

json2csv -i input.json -f date,temp_c,humidity --unwind weather_data > gnuplot_ready.csv

Notes #

• Replace weather_data with the actual name of your nested array in the JSON.
• Adjust the field list (date,temp_c,humidity) to match your data's keys.

Either approach will give you a clean CSV that Gnuplot can read easily. Just tweak the regex/field paths to match your specific JSON structure!

内容的提问来源于stack exchange，提问作者user277387