add a read_hdf() method that doesn't require pytables

2 files changed, 17 insertions, 2 deletions

diff --git a/utils/sddm/__init__.py b/utils/sddm/__init__.py
index 1f926df..d3a1be4 100644
--- a/utils/sddm/__init__.py
+++ b/utils/sddm/__init__.py
@@ -2,6 +2,8 @@ from __future__ import print_function, division
 import sys
 from itertools import izip_longest
 import os
+import h5py
+import pandas as pd
 
 IDP_E_MINUS  =    20
 IDP_MU_MINUS =    22
@@ -172,3 +174,12 @@ def setup_matplotlib(save=False):
 
         # Make the defalt font bigger
         matplotlib.rcParams['font.size'] = 12
+
+def read_hdf(filename, df):
+    """
+    Similar to pd.read_hdf() but doesn't require pytables. I'm adding this
+    function since the open science grid login nodes don't have pytables.
+    """
+    with h5py.File(filename) as f:
+        data = f[df][:]
+    return pd.DataFrame.from_records(data)
diff --git a/utils/submit-grid-jobs b/utils/submit-grid-jobs
index d4b1ae2..70bc4f9 100755
--- a/utils/submit-grid-jobs
+++ b/utils/submit-grid-jobs
@@ -439,6 +439,7 @@ if __name__ == '__main__':
     from sddm.plot_energy import prompt_event, gtid_sort, unwrap_50_mhz_clock
     import pandas as pd
     from sddm.dc import DC_MUON, DC_JUNK, DC_CRATE_ISOTROPY, DC_QVNHIT, DC_NECK, DC_FLASHER, DC_ESUM, DC_OWL, DC_OWL_TRIGGER, DC_FTS, DC_ITC, DC_BREAKDOWN
+    from sddm import read_hdf
 
     parser = argparse.ArgumentParser("submit grid jobs", formatter_class=argparse.ArgumentDefaultsHelpFormatter)
     parser.add_argument("filenames", nargs='*', help="input files")
@@ -550,8 +551,11 @@ if __name__ == '__main__':
             else:
                 check_call([zdab_cat,filename,"-o",output.name],stderr=f)
 
-        ev = pd.read_hdf(output.name, "ev")
-        rhdr = pd.read_hdf(output.name, "rhdr")
+        ev = read_hdf(output.name, "ev")
+        rhdr = read_hdf(output.name, "rhdr")
+
+        if len(ev) == 0:
+            continue
 
         if len(ev) and not args.reprocess and int(ev.iloc[0]['run']) in unique_runs:
             head, tail = split(filename)