Merge pull request #45 from mdoelker/master

Added discardUnmappedColumns option that throws away extra columns that ...

README.md

@@ -18,6 +18,7 @@ All methods accept the following `options`
 * `objectMode=true`: Ensure that `data` events have an object emitted rather than the stringified version set to false to have a stringified buffer.
 * `headers=false`: Ste to true if you expect the first line of your `CSV` to contain headers, alternatly you can specify an array of headers to use.
 * `ignoreEmpty=false`: If you wish to ignore empty rows.
+* `discardUnmappedColumns=false`: If you want to discard columns that do not map to a header.
 * `delimiter=','`: If your data uses an alternate delimiter such as `;` or `\t`.
    * **NOTE** When specifying an alternate `delimiter` you may only pass in a single character delimiter
 * `quote='"'`: The character to use to escape values that contain a delimiter.
@@ -473,7 +474,7 @@ var formatStream = csv
                 emailAddress: obj.Email_Address,
                 verified: obj.Verified
             };
-        });  
+        });
 csv
    .fromPath("in.csv", {headers: true})
    .pipe(formatStream)

lib/parser_stream.js

@@ -51,6 +51,7 @@ function ParserStream(options) {
     this.parser = createParser(options);
     this._headers = options.headers;
     this._ignoreEmpty = options.ignoreEmpty;
+    this._discardUnmappedColumns = options.discardUnmappedColumns;
     this.__objectMode = options.objectMode;
     this.__buffered = [];
     return this;
@@ -111,6 +112,7 @@ extended(ParserStream).extend({
             rows = data.rows;
             if (!this._parsedHeaders) {
                 var headers = this._headers;
+                var discardUnmappedColumns = this._discardUnmappedColumns;
                 if (extended.isBoolean(headers) && headers) {
                     headers = this.__handleLine(rows.shift(), 0, true);
                 }
@@ -120,7 +122,11 @@ extended(ParserStream).extend({
                     this.__transform = function (data, index) {
                         var ret = {}, i = -1, val;
                         if (data.length > headersLength) {
-                            self.emit("error", new Error("Unexpected Error: column header mismatch expected: " + headersLength + " columns got: " + data.length));
+                            if (discardUnmappedColumns) {
+                                data.splice(headersLength);
+                            } else {
+                                self.emit("error", new Error("Unexpected Error: column header mismatch expected: " + headersLength + " columns got: " + data.length));
+                            }
                         }
                         while (++i < headersLength) {
                             val = data[i];
@@ -268,4 +274,4 @@ extended(ParserStream).extend({
 
 });
 
-module.exports = ParserStream;
+module.exports = ParserStream;

test/assets/test23.csv

@@ -0,0 +1,4 @@
+first_name,last_name,email_address
+First1,Last1,[email protected]
+First2,Last2,[email protected],Unmapped1
+First3,Last3,[email protected],Unmapped1,Unmapped2

test/fast-csv.test.js

@@ -141,8 +141,14 @@ var expected14 = [
 
 var expected21 = [
     {"first_name": "First\n1", "last_name": "Last\n1", "email_address": "[email protected]", address: "1 Street St,\nState ST, 88888"},
-    {"first_name": "First\n2", "last_name": "Last\n2", "email_address": "[email protected]", address: "2 Street St,\nState ST, 88888"},
-]
+    {"first_name": "First\n2", "last_name": "Last\n2", "email_address": "[email protected]", address: "2 Street St,\nState ST, 88888"}
+];
+
+var expected23 = [
+    {"first_name": "First1", "last_name": "Last1", "email_address": "[email protected]"},
+    {"first_name": "First2", "last_name": "Last2", "email_address": "[email protected]"},
+    {"first_name": "First3", "last_name": "Last3", "email_address": "[email protected]"}
+];
 
 it.describe("fast-csv", function (it) {
 
@@ -461,6 +467,21 @@ it.describe("fast-csv", function (it) {
             });
     });
 
+    it.should("discard extra columns that do not map to a header with discardUnmappedColumns option", function (next) {
+        var actual = [];
+        csv
+            .fromPath(path.resolve(__dirname, "./assets/test23.csv"), {headers: true, discardUnmappedColumns: true})
+            .on("record", function (data, index) {
+                actual.push(data);
+            })
+            .on("error", next)
+            .on("end", function (count) {
+                assert.deepEqual(actual, expected23);
+                assert.equal(count, actual.length);
+                next();
+            });
+    });
+
     it.describe("alternate delimiters", function (it) {
         it.should("support tab delimiters", function (next) {
             var actual = [];
@@ -1000,4 +1021,4 @@ it.describe("fast-csv", function (it) {
                 });
         });
     });
-});
+});