问题
F#
I am interested to read a csv file and output a List< List< string > >
let readCsv (filepath:string) : string list list =
//.......................
input file:
Quote1,Quote2,Quote3
"Hello,World","He said:""Yes""",Example
Output:
// Type: string list list
[["Quote1";"Quote2";"Quote3"];
["Hello,World"; "He said:"Yes"";"Example"]]
Input2:
1,2,3,4,5,6
7,8,9,10,11,12
Output2:
// Type: string list list
[["1";"2";"3";"4";"5";"6"];
["7";"8";"9";"10";"11";"12"]]
However, some of the Nuget packages, e.g. CsvHelper, FileHelper, F#Data relies on defining a Class to "capture" the data, or defining a type by referring to a csv file.
https://joshclose.github.io/CsvHelper/
http://www.filehelpers.net/example/QuickStart/ReadWriteRecordByRecord/
http://fsharp.github.io/FSharp.Data/index.html
For example:
// In C#, from FileHelper Documentation
[DelimitedRecord(",")]
public class AbstractClass
{
public string Quote1;
public string Quote2;
public string Quote3;
}
or
// F# Data Documentation
type AbstractType = CsvProvider<"../example.csv">
But the input file may change in number of columns (and so I cannot define an abstract class)
Of course, I can just write regular expression to break up the input file line-by-line, but I am interested to know if someone else has already done it (or is it a standard library function).
Thank you.
回答1:
If you use FSharp.Data
there's a CsvFile
class which can read arbitrary CSV files.
e.g.
let csv = CsvFile.Load(filename, hasHeaders = true)
csv.Rows
|> Seq.map (fun r -> (r.["Image"], float r.["Size"]))
Would create a sequence of tuples from the "Image" and "Size" columns.
csv.Headers
is a string[] option
which contains the headers from the first line of the file.
let csv = CsvFile.Load(filename, hasHeaders = false)
csv.Rows
|> Seq.map (fun r -> r.Columns |> List.ofArray)
|> List.ofSeq
might be what you're after
来源:https://stackoverflow.com/questions/45412423/library-for-reading-csv-file-in-f