diff --git a/src/main/python/fallback_csv.py b/src/main/python/fallback_csv.py
index e5f4b89..47a3ed9 100644
--- a/src/main/python/fallback_csv.py
+++ b/src/main/python/fallback_csv.py
@@ -1,25 +1,58 @@
 import glob
 import os
 
-
-def read_file=
+def write_cachefile(cache_dir, data):
+    pass
 
 def generate(master_dir, from_time, to_time, cache_dir_fullpath):
 
-    timeframes = dict()
+    timeframes = []
 
     # read files
     files = glob.glob(master_dir + "/*.txt"):
     print(files)
 
-    for f in files:
-        timeframes
+    for fname in files:
+
+        start = None
+        end = None
+        data = []
+
+        # read file
+        with open(fname) as f:
+            first_line = True
+
+            # iterate through csv #
+            for line in f:
+
+                # skip header
+                if first_line:
+                    first_line = False
+                    continue
+
+                # read the line #
+                station_id, fulldate, dunno, temp, hum, dunno2 = line.split(";")
+
+                # parse date #
+                date = datetime.datetime.strptime(fulldate, "%y%m%d%H")
+
+                # set start and end #
+                if not start and date:
+                    start = date
+                elif date:
+                    end = date
+
+        # save values #
+        timeframes.append((start, end, data))
+
+    # find a fitting frame #
+    for start, end, data in timeframes:
+        if from_time >= start and to_time <= end:
+            write_cachefile(cache_dir_fullpath, data)
+            return
+
+    raise ValueError("Keine Datei mit passenden Daten gefunden. Bitte Readme lesen")
 
 
-    # glob files
-    # parse start end dates
-    # find a file that is a superset
-    # load the file
     # generate a correct cache file from it return
-    # fix the dates
     # if no suiteable file is found return an appropriate error