The Snap geoprocessing tool was introduced in Version 10 as a way of snapping together vertices of existing features. However, the function of this tool differs from what might be expected from manipulating features in an edit session.
The Snap geoprocessing tool belongs to the Editing Tools toolbox and can be run during an edit session, although an edit session is not required. The tool does not operate as if a feature is being moved manually in Data View. Instead, the behavior is similar to other geoprocessing tools that operate within a specified tolerance.
The effect of a snap is to move feature vertices that are within a certain distance of target features, so that input and target vertices are made coincident. This can result in 'stretching' or 'shrinking' of the dimensions of the features being snapped.
The vertices of the input feature are snapped to the nearest vertex, edge, or end point of the target dataset(s) within the specified distance. The distance, snap type (vertex, edge, or end), and target layer are input parameters to the Snap tool. Multiple snap types, each with its own snap distance, can be specified in one call to the tool.
In the special case where features of a layer are snapped to other features of the same layer, the feature with the lower Object ID or Feature ID will usually be snapped to the feature with the higher Object ID. However, the lower-to-higher precedence is not guaranteed.
In order to arrange the feature Object IDs in the desired order, use the geoprocessing Sort tool, also new with Version 10 in the Data Management Tools toolbox > General toolset. This tool reorders, in ascending or descending order, records in a feature class or table based on field values. Multiple sort fields can be specified. The reordered result is exported to a new dataset; it does not update the original source layer.
Refer to the Help documentation in the Related Links section for additional details of the tool behavior.