Clustering or partitioning is done in Wireless Sensor Networks to increase the lifetime of nodes by balancing the load and energy consumption in the network. Direct communication of nodes with the sink is an energy-consuming task. So, the nodes communicate with the sink using multi-hop communication. It has been observed that the formation of clusters provides a better solution to the multi-hop communication problem. Here, every member node in a cluster sends the messages to its Cluster Head (CH) as its next hop. The CHs collect data from all nodes in their clusters, aggregate and compress the data and forward it to the sink. If the sink is in direct communication range of the CH, the data can be sent directly, else a multi-hop route can be selected which consists of other CHs. The energy consumption of CH is more than the common nodes because of the extra computational work it must perform. Periodic CH selection is done to maintain the energy balance in the network. Partitioning the Region of Interest (RoI) into clusters, selecting the CH for each cluster, and choosing the communication method from CH to sink constitute the basic clustering or partitioning problem. The paper focuses on the comparative study of available techniques for clustering and aims to identify the drawbacks in them with the objective of outlining the future challenges to be overcome in the efficient clustering of WSNs.