There are heated debates on the existence of the global warming slowdown during the early 21^st century. Although efforts have been made to clarify or reconcile the controversy over the issue, it is not explicitly addressed, restricting the understanding of global temperature change particularly under the background of increasing greenhouse-gas concentrations. Here, using extensive temperature datasets, we comprehensively reexamine the existence of the slowdown under all existing definitions during all decadal-scale periods spanning 1990-2017. Results show that the short-term linear-trend dependent definitions of slowdown make its identification severely suffer from the period selection bias, which largely explains the controversy over its existence. Also, the controversy is further aggravated by the significant impacts of the differences between various datasets on the recent temperature trend and the different baselines for measuring slowdown prescribed by various definitions. However, when the focus is shifted from specific periods to the probability of slowdown events, we find the probability is significantly higher in the 2000s than in the 1990s, regardless of which definition and dataset are adopted. This supports a slowdown during the early 21^st century relative to the warming surge in the late 20^th century, despite higher greenhouse-gas concentrations. Furthermore, we demonstrate that this decadal-scale slowdown is not incompatible with the centennial-scale anthropogenic warming trend, which has been accelerating since 1850 and never pauses or slows. This work partly reconciles the controversy over the existence of the warming slowdown and the discrepancy between the slowdown and anthropogenic warming.