While running over the Run14 P18ih picoDsts, I believe I may have stumbled on some swapped tower IDs which are not accounted for in the BEMC mapping. The towerIDs should be corrected for when writing out the picoDsts, and I believe they are for all of the ones found on the Run 14 BEMC map, found here. However these towers seemed to be swapped don't show up on the run 14 BEMC mapping.  I do not know if this is for only run14 or if it affects other runs.

I suspect that the tower swapping goes as:
   3425, 3426, 3427, 3428 <----->   3433, 3434, 3435, 3436
   3445, 3446, 3447, 3448 <-----> *3452, 3454, 3455, 3456
  
*Note 3452 is marked as swapped with 3453 in the BEMC map. The towerIDs I am using are as seen by the analyzer.

This sort of swap seems bizarre to me, as most swaps in the BEMC map seem to be close range. Below is what suggests to me that they might be swapped

The following histograms are made for the towers I believe which are swapped. When the selected tower has a signal of ET > 2.0 GeV (after hadronic correction), I fill in the relative location of all charged tracks which pass basic QA cuts with pT > 3.0 GeV.  All of the plots have some noise, but the towers I found have a clear enhancement at eta~ +-0.4

Note tower 3428 made it onto my bad tower list so it has no entries but by the pattern I believe it to bad swapped with 3436.

 

---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Below is a snippet from the bemc map in which I highlighted in pink the towers above. The horizontal direction is changing in eta.

After finding these 16 towers, I searched for more towers which might be swapped, and I found 4 more pairs,

I believe 2821, 2822, 2823, 2824 <--------> 2801, 2802,  2803, 2804

These are adjacent in phi, which seems like a much more common swapping. Here are two example plots which leads me to belive they are swapped.

  swapped with

These aren't as convincing as the towers which are further separated in eta, but still suggest they are swapped.
Here again is a snippet from the bemc map in which I highlighted the additional 8 towers.

For completeness, below is an example of a histogram for a tower which is not swapped at the level of the analyzer. There is still noise but no enhancement away from (0,0). Also note this is after my hadronic correction. These plots are bootstrapped from my analysis, and I plan on making cleaner plots soon.