I may not understand your question, but if you are truly using an Xbar chart, a point on that chart correlates to multiple data points that make up the subgroup (and depends on the subgroup size how many points there are). Varying subgroup sizes across the chart is not recommended by any practitioner. This is because of how control charts work. The range charts are used to determine if the within subgroup variation is consistent/stable. The within subgroup variation is a function of the sources of variation that change within subgroup. If you change the subgroup sizes, you are effectively changing what sources of variation are captured within subgroup. The charts are no longer useful. The Xbar chart is a comparison chart. It compares the variation captured with subgroup (control limits) to the sources of variation captured between subgroup (plotted averages). Of course, it doesn't make any sense to do comparisons when the basis of comparison is not stable. (range charts depicting a stable process). Similarly, if you change your strategy for how the between subgroup samples are obtained, you change the sources of variation in the study and what comparisons are being made.
Please don't be offended, but may I suggest you research control chart theory before using the methodology. Start with Wheeler's book on Statistical Process Control (it's a bit of an easier read than Shewhart's book Economic Considerations of the Quality of Manufactured Product)
"All models are wrong, some are useful" G.E.P. Box
