make loadl less picky when it can’t merge an observable

python/loadleveller/mcextract.py

@@ -74,4 +74,9 @@ class MCArchive:
         selection.mean = orig.mean[mask,:]
         selection.error = orig.error[mask,:]
 
+        # if it is not a vector observable we can make it simpler for the user
+        if selection.mean.shape[-1] == 1:
+            selection.mean = selection.mean.flatten()
+            selection.error = selection.error.flatten()
+
         return selection

src/evalable.cpp

@@ -2,6 +2,7 @@
 #include "measurements.h"
 #include <fmt/format.h>
 #include <map>
+#include <iostream>
 
 namespace loadl {
 
@@ -27,9 +28,8 @@ void evalable::jackknife(const results &res, observable_result &obs_res) const {
 	size_t bin_count = -1; // maximal value
 	for(const auto &obs_name : used_observables_) {
 		if(res.observables.count(obs_name) <= 0) {
-			throw std::runtime_error(
-			    fmt::format("evalable '{}': used observable '{}' not found in Monte Carlo results.",
-			                name_, obs_name));
+			std::cerr << fmt::format("Warning: evalable '{}': used observable '{}' not found in Monte Carlo results. Skipping...", name_, obs_name);
+			return;
 		}
 		const auto &obs = res.observables.at(obs_name);