Effect sizes on two-way ANOVA post-hoc comparisons

Hi all, I'm new here. I am a researcher but not a stats guru. I've got a question about appropriate effect size statistics following a two-way ANOVA.

I typically use partial eta squared for overall main effect size in a 2-way ANOVA, and then either Hedge's g or Cohen's d for a standardized effect size for post-hoc comparisons. However, my question pertains using pooled data in the event that one main effect (eg: condition) is not significant, but the second is (time). So for example, there is no main effect of condition (3 levels) but there IS a main effect of time (also 3 levels).

Can I "pool" the data across the 3 levels of condition (which are repeated measures btw, not 3 separate subject groups) since there was no main effect here, which therefore leaves 3 time points (pre, post1 and post2), and then use Hedge's g as an effect size of pre v post1 and also post1 vs post2?