BUG: handle rounding issue with histogram edges on float32 data

Following inequality causes wrong counting at the edges and can be avoided by
making the edge array of the same type as the input data.

In [1]: np.around(np.float64(6010.36962890625), 5)
Out[1]: 6010.3696300000001
In [2]: np.around(np.float32(6010.36962890625), 5)
Out[2]: 6010.3701

Closes gh-4799

1 files changed, 7 insertions, 2 deletions

diff --git a/numpy/lib/function_base.py b/numpy/lib/function_base.py
index 00bfab6ba..f625bcb90 100644
--- a/numpy/lib/function_base.py
+++ b/numpy/lib/function_base.py
@@ -336,6 +336,11 @@ def histogramdd(sample, bins=10, range=None, normed=False, weights=None):
             smin[i] = smin[i] - .5
             smax[i] = smax[i] + .5
 
+    # avoid rounding issues for comparisons when dealing with inexact types
+    if np.issubdtype(sample.dtype, np.inexact):
+        edge_dt = sample.dtype
+    else:
+        edge_dt = float
     # Create edge arrays
     for i in arange(D):
         if isscalar(bins[i]):
@@ -344,9 +349,9 @@ def histogramdd(sample, bins=10, range=None, normed=False, weights=None):
                     "Element at index %s in `bins` should be a positive "
                     "integer." % i)
             nbin[i] = bins[i] + 2  # +2 for outlier bins
-            edges[i] = linspace(smin[i], smax[i], nbin[i]-1)
+            edges[i] = linspace(smin[i], smax[i], nbin[i]-1, dtype=edge_dt)
         else:
-            edges[i] = asarray(bins[i], float)
+            edges[i] = asarray(bins[i], edge_dt)
             nbin[i] = len(edges[i]) + 1  # +1 for outlier bins
         dedges[i] = diff(edges[i])
         if np.any(np.asarray(dedges[i]) <= 0):