From 029535b102c484afb8bea2c0c524bc5189dc154f Mon Sep 17 00:00:00 2001
From: Patrick Peglar <patrick.peglar@metoffice.gov.uk>
Date: Mon, 17 May 2021 18:19:17 +0100
Subject: [PATCH] Remove workaround when dask-wrapping PP data, obsoleted by
 #4135.

---
 lib/iris/fileformats/pp.py | 30 ++++++++-----------
 lib/iris/util.py           | 61 --------------------------------------
 2 files changed, 12 insertions(+), 79 deletions(-)

diff --git a/lib/iris/fileformats/pp.py b/lib/iris/fileformats/pp.py
index 406da925b1..7589d27922 100644
--- a/lib/iris/fileformats/pp.py
+++ b/lib/iris/fileformats/pp.py
@@ -38,7 +38,6 @@
 )
 import iris.fileformats.rules
 import iris.coord_systems
-from iris.util import _array_slice_ifempty
 
 try:
     import mo_pack
@@ -594,23 +593,18 @@ def ndim(self):
         return len(self.shape)
 
     def __getitem__(self, keys):
-        # Check for 'empty' slicings, in which case don't fetch the data.
-        # Because, since Dask v2, 'dask.array.from_array' performs an empty
-        # slicing and we must not fetch the data at that time.
-        result = _array_slice_ifempty(keys, self.shape, self.dtype)
-        if result is None:
-            with open(self.path, "rb") as pp_file:
-                pp_file.seek(self.offset, os.SEEK_SET)
-                data_bytes = pp_file.read(self.data_len)
-                data = _data_bytes_to_shaped_array(
-                    data_bytes,
-                    self.lbpack,
-                    self.boundary_packing,
-                    self.shape,
-                    self.src_dtype,
-                    self.mdi,
-                )
-            result = data.__getitem__(keys)
+        with open(self.path, "rb") as pp_file:
+            pp_file.seek(self.offset, os.SEEK_SET)
+            data_bytes = pp_file.read(self.data_len)
+            data = _data_bytes_to_shaped_array(
+                data_bytes,
+                self.lbpack,
+                self.boundary_packing,
+                self.shape,
+                self.src_dtype,
+                self.mdi,
+            )
+        result = data.__getitem__(keys)
 
         return np.asanyarray(result, dtype=self.dtype)
 
diff --git a/lib/iris/util.py b/lib/iris/util.py
index d2fda82c0f..aea07a877c 100644
--- a/lib/iris/util.py
+++ b/lib/iris/util.py
@@ -960,67 +960,6 @@ def __lt__(self, other):
             return NotImplemented
 
 
-def _array_slice_ifempty(keys, shape, dtype):
-    """
-    Detect cases where an array slice will contain no data, as it contains a
-    zero-length dimension, and produce an equivalent result for those cases.
-
-    The function indicates 'empty' slicing cases, by returning an array equal
-    to the slice result in those cases.
-
-    Args:
-
-    * keys (indexing key, or tuple of keys):
-        The argument from an array __getitem__ call.
-        Only tuples of integers and slices are supported, in particular no
-        newaxis, ellipsis or array keys.
-        These are the types of array access usage we expect from Dask.
-    * shape (tuple of int):
-        The shape of the array being indexed.
-    * dtype (numpy.dtype):
-        The dtype of the array being indexed.
-
-    Returns:
-        result (np.ndarray or None):
-            If 'keys' contains a slice(0, 0), this is an ndarray of the correct
-            resulting shape and provided dtype.
-            Otherwise it is None.
-
-    .. note::
-
-        This is used to prevent DataProxy arraylike objects from fetching their
-        file data when wrapped as Dask arrays.
-        This is because, for Dask >= 2.0, the "dask.array.from_array" call
-        performs a fetch like [0:0, 0:0, ...], to 'snapshot' array metadata.
-        This function enables us to avoid triggering a file data fetch in those
-        cases :  This is consistent because the result will not contain any
-        actual data content.
-
-    """
-    # Convert a single key into a 1-tuple, so we always have a tuple of keys.
-    if isinstance(keys, tuple):
-        keys_tuple = keys
-    else:
-        keys_tuple = (keys,)
-
-    if any(key == slice(0, 0) for key in keys_tuple):
-        # An 'empty' slice is present :  Return a 'fake' array instead.
-        target_shape = list(shape)
-        for i_dim, key in enumerate(keys_tuple):
-            if key == slice(0, 0):
-                # Reduce dims with empty slicing to length 0.
-                target_shape[i_dim] = 0
-        # Create a prototype result : no memory usage, as some dims are 0.
-        result = np.zeros(target_shape, dtype=dtype)
-        # Index with original keys to produce the desired result shape.
-        # Note : also ok in 0-length dims, as the slice is always '0:0'.
-        result = result[keys]
-    else:
-        result = None
-
-    return result
-
-
 def create_temp_filename(suffix=""):
     """Return a temporary file name.