问题
String[] values = line.split(",");
Long locId = Long.parseLong(replaceQuotes(values[0]));
String country = replaceQuotes(values[1]);
String region = replaceQuotes(values[2]);
String city = replaceQuotes(values[3]);
String postalCode = replaceQuotes(values[4]);
String latitude = replaceQuotes(values[5]);
String longitude = replaceQuotes(values[6]);
String metroCode = replaceQuotes(values[7]);
String areaCode = replaceQuotes(values[8]);
//...
public String replaceQuotes(String txt){
txt = txt.replaceAll("\"", "");
return txt;
}
I'm using the code above to parse a CSV with data in this format:
828,"US","IL","Melrose Park","60160",41.9050,-87.8641,602,708
However, when I encounter a line of data such as the following I get java.lang.ArrayIndexOutOfBoundsException: 7
1,"O1","","","",0.0000,0.0000,,
Does this mean that any time I even try to access the value at values[7]
, an Exception will be thrown?
If so, how do I parse lines that don't contain data in that position of the text line?
回答1:
First of all, String.split()
is not a great CSV parser: it doesn't know about quotes and will mess up as soon as one of your quoted values contains a comma.
That being said, by default String.split() leaves out empty trailing elements. You can influence that by using the two-argument variant:
String[] values = line.split(",", -1);
-1
(or any negative value) means that the array will be as large as necessary.- Using a positive value gives a maximum amount of splits to be done (meaning that everything beyond that will be a single value, even if it contains a comma).
0
(the default if you use the one-argument value) means that the array will be as large as necessary, but empty trailing values will be left out of the array (exactly as it happens to you).
回答2:
As a general rule you should never, ever hack up your own (faulty) parser if a working one already exists. CSV is not easy to parse correctly, and String.split will not do the job since CSV allows ,
to be used between "
's without working as separaters.
Consider using OpenCSV. This will solve both the problem you have now and the problem you will face when a user uses a ,
as part of the data.
来源:https://stackoverflow.com/questions/6581932/how-to-avoid-triggering-an-arrayindexoutofboundsexception-while-parsing-empty-po