Similar to the case of condensed IP-tables, the insertion and deletion algorithms have to be changed when using endpoint IP-tables. However, it is possible to accurately maintain the set V''(R) of timepoints within an IP-table.
Figure 7.14 shows the actions that have to be performed when a tuple r is inserted into the temporal relation R. If r's endpoint r.te is not contained in V''(R) then it is added. Consequently, there is always a such that t''je = r.te when it comes to the stage of modifying the s[j] and o[j] values. In contrast to je, the index js is determined as in the case of a condensed IP-table by looking for the nearest such that . The modification of the and arrays works as for complete and condensed IP-tables.
The deletion algorithm for endpoint IP-tables is straightforward. It is shown in figure 7.15. It determines js and je as in the case of insertion, then modifies the and arrays before finally checking whether the r.te = t''je is the endpoint of an interval other than r. If it is not then it can be removed from the set V''(R). This removal is not trivial as the value of s[je] has to be incorporated into the one of the timepoint t''je+1 which follows t''je within the ordered set V''(R).