K-means algorithm implementation

In my data set contains more than 10 attributes how to find k means for that 10 attributes or we have to choose attributes our self like 4 out of 10 attributes

You can (and should) use all of the features.