GNUPLOT技术问题：如何在循环中用CSV表头作为绘图标题

Got it, let's get those headers showing up as your histogram titles! Here's a step-by-step fix that works with your existing setup, plus some tweaks to make it more flexible:

1. Fix the Data Separator First #

Your CSV uses semicolons (;) as separators, so your initial line needs to be corrected to:

set datafile separator ";"

This ensures gnuplot correctly parses each column instead of treating the whole line as a single value.

2. Extract and Store the Headers #

We'll read the first line of your CSV into an array, so we can reference each header when plotting its column:

# Read the header line from your CSV (works on Linux/macOS; see note for Windows)
header_line = system("head -n 1 your_data.csv")

# Create an array to hold individual headers (adjust the size if you have more than 100 columns)
array column_headers[100]

# Split the header line into separate entries
do for [i=1:words(header_line)] {
    column_headers[i] = word(header_line, i)
}

Note for Windows users: Replace the system() line with this to extract the first line:

header_line = system("type your_data.csv | findstr /n "^" | findstr "^1:" | for /f "tokens=2 delims=:" %a in ('more') do @echo %a")

3. Auto-Detect Column Count (Optional But Handy) #

Instead of hardcoding the number of columns, let gnuplot count them automatically:

# Get total columns from the CSV
stats 'your_data.csv' using 0 nooutput
total_columns = STATS_columns

4. Plot Each Column with Its Header as Title #

Now loop through every column, set the title to the matching header, and save each plot to a unique file:

# Set your histogram style preferences (adjust to match your original setup)
set terminal pngcairo enhanced font 'Arial,12'
set style fill solid 0.8
set binwidth 2
set xlabel "Value Bins"
set ylabel "Frequency"

# Loop through each column to generate histograms
do for [col=1:total_columns] {
    # Set plot title to the corresponding header
    set title sprintf("Histogram: %s", column_headers[col])
    
    # Save to a unique file (e.g., histogram_headerName.png or column number)
    set output sprintf("histogram_%s.png", column_headers[col])
    
    # Plot the frequency histogram for the column
    plot 'your_data.csv' using (bin($col, binwidth)) smooth freq title column_headers[col] with boxes
}

Key Notes: #

• The bin($col, binwidth) function groups your numeric data into bins of the size you specified (2 in your case).
• smooth freq calculates how many data points fall into each bin, which is exactly what you need for a histogram.
• Using sprintf() lets you create dynamic filenames and titles based on the header text, so your output files are easy to identify.

If you prefer to stick with your original set style data histogram setup, just replace the plot line with:

plot 'your_data.csv' using col title column_headers[col]

But for numeric data histograms, the bin() + smooth freq method is usually more accurate.

内容的提问来源于stack exchange，提问作者WolfiG