Grouping objects by multiple columns with Lodash o

问题:

I have following object records:

 {  
   "notes":[  
      {  
         "id":1,
         "description":"hey",
         "userId":2,
         "replyToId":null,
         "postId":2,
         "parentId":null
      },
      {  
         "id":5,
         "description":"hey test",
         "userId":3,
         "replyToId":null,
         "postId":2,
         "parentId":null
      },
      {  
         "id":2,
         "description":"how are you",
         "userId":null,
         "replyToId":2,
         "postId":2,
         "parentId":null,
         "user":null
      }
   ]
}

I want to output it as:

2 
  object with id 1
  object with id 2 (because replyToId value is same as userId
3
  object with id 5

So basically I want to consider UserId and replyToId value under the same group.

I have build my own mixin under lodash, wrapping groupBy method as:

mixin({
    splitGroupBy: function(list, groupByIter){
        if (_.isArray(groupByIter)) {
            function groupBy(obj) {
                return _.forEach(groupByIter, function (key){
                    if ( !!obj[key] ) return obj[key]
                });

            }
        } else {
            var groupBy = groupByIter;
        }

        debugger;

        var groups = _.groupBy(list, groupBy);

        return groups;
    }
});

Call looks like this:

_.splitGroupBy(data.notes,['userId', 'replyToId']);

The output is coming without group. Even when I have tried with _.map instead _.forEach the split is not happening correctly.

回答1:

A solution using underscore:

    var props = ['userId', 'replyToId'];

    var notNull = _.negate(_.isNull);

    var groups = _.groupBy(record.notes, function(note){
        return _.find(_.pick(note, props), notNull);
    });

回答2:

This can probably done much prettier, but it should work:

lodash.mixin({
  splitGroupBy: function(list, groupByIter) {
    var _ = this, groupBy;
    if (lodash.isArray(groupByIter)) {
      groupBy = function(obj) {
        return _(obj) .pick(groupByIter)
                      .values()
                      .without(null, undefined)
                      .first();
      };
    } else {
      groupBy = groupByIter;
    }
    var groups = _.groupBy(list, groupBy);
    return groups;
  }
});

回答3:

You could map your list of attributes to their respective values and pick the first non falsy value as your group key:

_.mixin({
    splitGroupBy: function(list, groupByIter){
        if (!_.isArray(groupByIter))
            return _.groupBy(list, groupByIter);

        return _.groupBy(list, function(o) {
            var values = _.map(groupByIter, function(k) {
                return o[k];
            });
            return _.find(values);
        });
    }
});

var data = {  
   "notes":[  
      {  
         "id":1,
         "userId":2,
         "replyToId":null
      },
      {  
         "id":5,
         "userId":3,
         "replyToId":null
      },
      {  
         "id":2,
         "userId":null,
         "replyToId":2
      }
   ]
};

_.mixin({
    splitGroupBy: function(list, groupByIter){
        if (!_.isArray(groupByIter))
            return _.groupBy(list, groupByIter);

        return _.groupBy(list, function(o) {
            var values = _.map(groupByIter, function(k) {
                return o[k];
            });
            return _.find(values);
        });
    }
});

snippet.log(JSON.stringify(_.splitGroupBy(data.notes,['userId', 'replyToId'])));

<script src="https://cdnjs.cloudflare.com/ajax/libs/underscore.js/1.8.3/underscore-min.js"></script>

<!-- Provides the `snippet` object, see http://meta.stackexchange.com/a/242144/134069 -->
<script src="http://tjcrowder.github.io/simple-snippets-console/snippet.js"></script>

回答4:

Assuming userId and replyToId are mutually exclusive (i.e. you either have a userId or a replyToId, but never both) as they are in the sample data, then specifying a custom grouping function works:

_.groupBy(data.notes, function(note) {
    return note.userId || note.replyToId;
});