From e41b03ca574ed79c58bad9dc9e72d84dc98af19b Mon Sep 17 00:00:00 2001
From: Lionel GUEZ <guez@lmd.ipsl.fr>
Date: Fri, 11 Oct 2024 20:19:31 +0200
Subject: [PATCH 1/3] Add option `my_range` to method iterRecords

Using iterRecords with a range option should be faster than calling
record within a loop, since we avoid the multiple calls to seek.
---
 shapefile.py | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/shapefile.py b/shapefile.py
index 12af74d..07bc3c0 100644
--- a/shapefile.py
+++ b/shapefile.py
@@ -1809,7 +1809,7 @@ def records(self, fields=None):
                 records.append(r)
         return records
 
-    def iterRecords(self, fields=None):
+    def iterRecords(self, fields=None, my_range=None):
         """Returns a generator of records in a dbf file.
         Useful for large shapefiles or dbf files.
         To only read some of the fields, specify the 'fields' arg as a
@@ -1820,7 +1820,9 @@ def iterRecords(self, fields=None):
         f = self.__getFileObj(self.dbf)
         f.seek(self.__dbfHdrLength)
         fieldTuples, recLookup, recStruct = self.__recordFields(fields)
-        for i in xrange(self.numRecords):
+        if my_range is None:
+            my_range = xrange(self.numRecords)
+        for i in my_range:
             r = self.__record(
                 oid=i, fieldTuples=fieldTuples, recLookup=recLookup, recStruct=recStruct
             )

From 4efef9fdb5a1d188a33316793cb964d33daa63df Mon Sep 17 00:00:00 2001
From: Lionel GUEZ <guez@lmd.ipsl.fr>
Date: Tue, 15 Oct 2024 06:57:12 +0200
Subject: [PATCH 2/3] Revert "Add option `my_range` to method iterRecords"

This reverts commit e41b03ca. JamesParrott pointed that I did not understand
the way `__record` works: __record does not use oid to find the correct
record, it just assumes it is the correct oid for the current position.
---
 shapefile.py | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/shapefile.py b/shapefile.py
index 07bc3c0..12af74d 100644
--- a/shapefile.py
+++ b/shapefile.py
@@ -1809,7 +1809,7 @@ def records(self, fields=None):
                 records.append(r)
         return records
 
-    def iterRecords(self, fields=None, my_range=None):
+    def iterRecords(self, fields=None):
         """Returns a generator of records in a dbf file.
         Useful for large shapefiles or dbf files.
         To only read some of the fields, specify the 'fields' arg as a
@@ -1820,9 +1820,7 @@ def iterRecords(self, fields=None, my_range=None):
         f = self.__getFileObj(self.dbf)
         f.seek(self.__dbfHdrLength)
         fieldTuples, recLookup, recStruct = self.__recordFields(fields)
-        if my_range is None:
-            my_range = xrange(self.numRecords)
-        for i in my_range:
+        for i in xrange(self.numRecords):
             r = self.__record(
                 oid=i, fieldTuples=fieldTuples, recLookup=recLookup, recStruct=recStruct
             )

From 811d32909e0b4889efa2357dd32682e324222225 Mon Sep 17 00:00:00 2001
From: Lionel GUEZ <guez@lmd.ipsl.fr>
Date: Tue, 15 Oct 2024 08:02:03 +0200
Subject: [PATCH 3/3] Add method `iterRecords_range`

Using the method `iterRecords_range` should be somewhat faster than
calling the method `record` within a loop, since we avoid the repeated
calls to seek inside `record`.
---
 shapefile.py | 26 ++++++++++++++++++++++++++
 1 file changed, 26 insertions(+)

diff --git a/shapefile.py b/shapefile.py
index 12af74d..906cdda 100644
--- a/shapefile.py
+++ b/shapefile.py
@@ -1827,6 +1827,32 @@ def iterRecords(self, fields=None):
             if r:
                 yield r
 
+    def iterRecords_range(self, start, stop, fields=None):
+        """Returns a generator of records in a dbf file, for a range
+        of oid.  Useful for large shapefiles or dbf files.  To only
+        read some of the fields, specify the 'fields' arg as a list of
+        one or more fieldnames.
+
+        """
+        if self.numRecords is None:
+            self.__dbfHeader()
+        f = self.__getFileObj(self.dbf)
+        start = self.__restrictIndex(start)
+        if abs(stop) > self.numRecords:
+            raise IndexError("Record index out of range.")
+        if stop < 0:
+            stop = range(self.numRecords)[stop]
+        recSize = self.__recordLength
+        f.seek(0)
+        f.seek(self.__dbfHdrLength + (start * recSize))
+        fieldTuples, recLookup, recStruct = self.__recordFields(fields)
+        for i in xrange(start, stop):
+            r = self.__record(
+                oid=i, fieldTuples=fieldTuples, recLookup=recLookup, recStruct=recStruct
+            )
+            if r:
+                yield r
+
     def shapeRecord(self, i=0, fields=None, bbox=None):
         """Returns a combination geometry and attribute record for the
         supplied record index.