The “active” selection? about d3-selection HOT 1 CLOSED

I recently changed the selection function arguments to {d, i, nodes}, partly to mimic standard JavaScript array methods that take {element, i, array}. This provides another way to access the current element: nodes[i], although perhaps not as convenient as the proposed d3.node. (The group can also be used to access the parent, as nodes._parent, although that requires the semi-private API.)

Also, now that selections are back to being bilevel rather than arbitrary hierarchies, the need to reconstruct the active selection is less important. Yes, d3.select(this) within selection.each loses the index and parent from the originating selection, but in most cases you don’t care about this information anyway. It feels like bilevel selections encourage you to think “flat” (for example, pushing data down to children as needed), rather than hierarchical selections which encourage more internal (inherited) state.

I think d3.activeTransition should require a node, rather than infer it from the “active” node. It feels a little too magical otherwise.

from d3-selection.