Octave代码运行缓慢需优化：循环处理文件提速方案咨询

Hey there! I see your current loop-based approach for processing that text file is running too slow—let's fix that. The main issue with your code is that manual looping and repeated string splitting adds a lot of overhead, especially with larger files. Octave has built-in functions optimized for exactly this kind of numerical text data, so we can replace that entire block with something way more efficient.

Why Your Original Code Is Slow #

Your current workflow does three steps that add unnecessary overhead:

• Reads the entire file into a single string, then splits it into lines
• Loops through each line to split into individual string cells
• Uses cellfun to convert each cell to a number
  Each of these steps involves iterating over elements in Octave, which is much slower than using vectorized/optimized built-in functions.

Faster Alternatives #

1. Use dlmread (Simplest & Fastest) #

Since your data is space-separated numerical values, dlmread is perfect—it directly reads the file into a numerical matrix in one line:

f = dlmread("C:/Backup/Thermo Backups/2018-05-17.txt");

This function skips all the manual string splitting and conversion steps entirely, as it's designed specifically for delimited numerical data.

2. Use textscan (More Flexible) #

If you ever need to handle more complex formatting (like mixed data types or specific column formats), textscan is a great option. It's still way faster than your loop:

# Open the file for reading
fid = fopen("C:/Backup/Thermo Backups/2018-05-17.txt", "r");
# Read all columns as floats, collect into a single matrix
data = textscan(fid, "%f %f %f %f %f %f %f %f %f", "CollectOutput", true);
# Close the file handle
fclose(fid);
# Extract the numerical matrix from the cell output
f = data{1};

The %f specifiers match your 9 columns of floating-point numbers, and CollectOutput combines them into one matrix instead of separate cells.

Quick Test #

For your sample data, both methods will produce the same numerical matrix as your original code, but in a fraction of the time—especially as your file size grows.

内容的提问来源于stack exchange，提问作者tomkut