Just use csv:from-file "/path/to/myfile.csv"
! See from-file for more information.
For really big files, you may not want to store the entire file in memory, but rather just process it a line at a time. For instance, if you want to sum each of the columns of a numeric CSV file, you can do:
to-report sum-columns [ file ]
file-open file
set result csv:from-row file-read-line
while [ not file-at-end? ] [
let row csv:from-row file-read-line
set result (map [ [ [col-total new-val] -> col-total + new-val] result row)
]
file-close
report result
end
You can also use this technique to…
Here’s an example model that reads in a file one line per tick:
globals [ data ]
to setup
clear-all
file-close-all % Close any files open from last run
file-open "data.csv"
% other setup goes here
reset-ticks
end
to go
if file-at-end? [ stop ]
set data csv:from-row file-read-line
% model update goes here
tick
end
Just use csv:to-file "/path/to/myfile.csv" my-data
! See to-file for more information.
Parses the given string as though it were a row from a CSV file and returns it as a list of values. For example:
observer> show csv:from-row "one,two,three"
observer: ["one" "two" "three"]
Quotes can be used when items contain commas:
observer> show csv:from-row "there's,a,comma,\"in,here\""
observer: ["there's" "a" "comma" "in,here"]
You can put two quotes in a row to put an actual quote in an entry. If the entry is not quoted, you can just use one quote:
observer> foreach (csv:from-row "he said \"hi there\",\"afterwards, she said \"\"hello\"\"\"") print
he said "hi there"
afterwards, she said "hello"
Number-like-entries will be parsed as numbers:
observer> show csv:from-row "1,-2.5,1e3"
observer: [1 -2.5 1000]
true
and false
with any capitalization will be parsed as booleans:
observer> show csv:from-row "true,TRUE,False,falsE"
observer: [true true false false]
To use a different delimiter, you can specify a second, optional argument. Only single character delimiters are supported:
observer> show (csv:from-row "1;2;3" ";")
observer: [1 2 3]
Different types of values can be mixed freely:
observer> show csv:from-row "one,2,true"
observer: ["one" 2 true]
Parses a string representation of one or more CSV rows and returns it as a list of lists of values. For example:
observer> show csv:from-string "1,two,3\nfour,5,true"
observer: [[1 "two" 3] ["four" 5 true]]
Parses an entire CSV file to a list of lists of values. For example, if we have a file example.csv
that contains:
1,2,3
4,5,6
7,8,9
10,11,12
Then, we get:
observer> show csv:from-file "example.csv"
observer: [[1 2 3] [4 5 6] [7 8 9] [10 11 12]]
The parser doesn’t care if the rows have different numbers of items on them. The number of items in the rows list
will always be <number of delimiters> + 1
, though blank lines are skipped. This makes handling files with headers
quite easy. For instance, if we have header.csv
that contains:
My Data
2/1/2015
Parameters:
start,stop,resolution,population,birth?
0,4,1,100,true
Data:
time,x,y
0,0,0
1,1,1
2,4,8
3,9,27
This gives:
observer> foreach csv:from-file "header.csv" show
observer: ["My Data"]
observer: ["2/1/2015"]
observer: ["Parameters:"]
observer: ["start" "stop" "resolution" "population" "birth?"]
observer: [0 4 1 100 true]
observer: ["Data:"]
observer: ["time" "x" "y"]
observer: [0 0 0]
observer: [1 1 1]
observer: [2 4 8]
observer: [3 9 27]
Reports the given list as a CSV row. For example:
observer> show csv:to-row ["one" 2 true]
observer: "one,2,true"
Reports the given list of lists as a CSV string. For example:
observer> show csv:to-string [[1 "two" 3] [4 5]]
observer: "1,two,3\n4,5"
Writes the given list of lists to a new CSV file. For example:
observer> csv:to-file "myfile.csv" [[1 "two" 3] [4 5]]
will result in a file myfile.csv
containing:
1,two,3
4,5