String Match for Plotting in Excel

2 visualizzazioni (ultimi 30 giorni)
Amanda
Amanda il 30 Apr 2013
I have a simple CSV Data file. See below. I need to extract data for each individual state with its temperature. For example, I need a script that can find the string New York and get the associated temperature data for New York, and put it in an array for plotting. So the output would be:
New_York = [83,55], Indiana = [70,60]
State Temperature (headerline)
New York 83
Indiana 70
Texas 72
California 80
Indiana 60
Texas 61
California 92
New York 55
Thanks, Amanda

Risposta accettata

Cedric
Cedric il 30 Apr 2013
Modificato: Cedric il 30 Apr 2013
If states had no space in their names, or if you had commas as delimiters in the CSV file, you could go for a variant of
[state, temp] = textread('myFile.csv', '%s %d', 'delimiter', ',', ...
'headerlines', 1) ;
Now as it seems that there are spaces in names and no comma as delimiter, you can read line by line and extract states and temperatures with more specific functions (TEXTSCAN, FSCANF, FGETL+SSCANF, etc) and based on position if needed (e.g. start reading temperatures from char. 15 on). Then you can use STRCMPI to find indices of relevant states, and get corresponding temperatures based on these indices. But you could also go for a solution based on regular expressions (less common approach for this kind of structured data), that I illustrate below:
>> buffer = fileread('myFile.csv') ;
>> state = 'New York' ;
>> temp = str2double(regexpi(buffer, sprintf('(?<=%s\\s*)\\d*', state), ...
'match'))
temp =
83 55
>> state = 'California' ;
>> temp = str2double(regexpi(buffer, sprintf('(?<=%s\\s*)\\d*', state), ...
'match'))
temp =
80 92
Note that I wrote this in a concise manner, but we do the following in fact:
>> pattern = sprintf('(?<=%s\\s*)\\d*', state) ;
>> match = regexpi(buffer, pattern, 'match') ;
>> temp = str2double(match) ;
If you look at the pattern for New Work:
>> pattern
pattern =
(?<=New York\s*)\d*
It tells regexp to match
  • As many numeric characters as possible: \d*
  • Preceded by (positive look behind (?<=)) the literal New York followed by as many white spaces as possible: New York\s*
  3 Commenti
Cedric
Cedric il 30 Apr 2013
You're welcome! I updated my answer so you have a bit more explanations about REGEXP.
Amanda
Amanda il 30 Apr 2013
Thanks for the more detail explanation. Some reason, data extraction from a file has always been a challenge.

Accedi per commentare.

Più risposte (0)

Categorie

Scopri di più su MATLAB in Help Center e File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by