Added: toDate() now supports formats.

My dataset has the same dates written in two styles: Nov-09 and 11/09
You can now parse this using: value.toDate2('MM/yy','MMM-yy').toString('yyyy-MM')



git-svn-id: http://google-refine.googlecode.com/svn/trunk@414 7d457c2a-affb-35e4-300a-418c747d4874
This commit is contained in:
Will Moffat 2010-04-07 19:12:08 +00:00
parent 4171db6540
commit efb9a43b73

View File

@ -5,6 +5,10 @@ import java.util.Properties;
import org.json.JSONException; import org.json.JSONException;
import org.json.JSONWriter; import org.json.JSONWriter;
import java.text.SimpleDateFormat;
import java.util.Date;
import java.util.GregorianCalendar;
import com.metaweb.gridworks.expr.util.CalendarParser; import com.metaweb.gridworks.expr.util.CalendarParser;
import com.metaweb.gridworks.expr.util.CalendarParserException; import com.metaweb.gridworks.expr.util.CalendarParserException;
import com.metaweb.gridworks.gel.Function; import com.metaweb.gridworks.gel.Function;
@ -12,33 +16,61 @@ import com.metaweb.gridworks.gel.Function;
public class ToDate implements Function { public class ToDate implements Function {
public Object call(Properties bindings, Object[] args) { public Object call(Properties bindings, Object[] args) {
if (args.length == 1 || args.length == 2) { if (args.length == 0) {
Object o1 = args[0]; // missing value, can this happen?
if (o1 != null && o1 instanceof String) { }
boolean month_first = true; if (!(args[0] instanceof String)) {
if (args.length == 2) { // ignore cell values that aren't strings
Object o2 = args[1]; return null;
if (o2 != null && o2 instanceof Boolean) { }
month_first = ((Boolean) o2).booleanValue(); String o1 = (String) args[0];
}
// "o, boolean month_first (optional)"
if (args.length == 1 || (args.length == 2 && args[1] instanceof Boolean)) {
boolean month_first = true;
if (args.length == 2) {
month_first = (Boolean) args[1];
}
try {
return CalendarParser.parse( o1, (month_first) ? CalendarParser.MM_DD_YY : CalendarParser.DD_MM_YY);
} catch (CalendarParserException e) {
// do something about
}
}
// "o, format1, format2 (optional), ..."
if (args.length>=2) {
for (int i=1;i<args.length;i++) {
if (!(args[i] instanceof String)) {
// skip formats that aren't strings
continue;
} }
String format = (String) args[i];
SimpleDateFormat formatter = new SimpleDateFormat(format);
Date date = null;
try { try {
return CalendarParser.parse((String) o1, (month_first) ? CalendarParser.MM_DD_YY : CalendarParser.DD_MM_YY); date = formatter.parse(o1);
} catch (CalendarParserException e) { } catch (java.text.ParseException e) {
// do something about // ignore
}
if (date != null) {
GregorianCalendar c = new GregorianCalendar();
c.setTime(date);
return c;
} }
} }
} }
return null; return null;
} }
public void write(JSONWriter writer, Properties options) public void write(JSONWriter writer, Properties options)
throws JSONException { throws JSONException {
writer.object(); writer.object();
writer.key("description"); writer.value("Returns o converted to a date object"); writer.key("description"); writer.value("Returns o converted to a date object, you can hint if the day or the month is listed first, or give an ordered list of possible formats using this syntax: http://java.sun.com/j2se/1.4.2/docs/api/java/text/SimpleDateFormat.html");
writer.key("params"); writer.value("o, boolean month_first (optional)"); writer.key("params"); writer.value("o, boolean month_first / format1, format2, ... (all optional)");
writer.key("returns"); writer.value("date"); writer.key("returns"); writer.value("date");
writer.endObject(); writer.endObject();
} }