Or is there any other way to represent a huge data table except by pagination or a [load more] button on a scrolling table?
A third approach might be to provide a way to further filter that result set, especially if the search is parametric. UIE has a good intro on faceted search:
Faceted search lets users refine or navigate a collection of information by using a number of discrete attributes – the so-called facets. A facet represents a specific perspective on content that is typically clearly bounded and mutually exclusive
If you can, dynamically analyse the result set to determine what are relevant facets or facet values (i.e. exclude any facet values wherein selection would result in zero results). Then, the user can drill down into the initial result set and triangulate a smaller, more relevant, result set.
For example, if I search for 'Hem' on LinkedIN I get 27,279 results. Far more than I could possibly browse (and LI actually limits pagination at 500 results). However, as you can see LI also offers to let me refine my search by presenting a number of search facets. I could, for example, reduce that 27,279 result set to just 2,543 persons by specifying "India" as the location. I could further refine by specifying a specific Industry or Company.