Add a tool to merge several podio files into a single one

tools/CMakeLists.txt

@@ -3,6 +3,7 @@ install(PROGRAMS ${CMAKE_CURRENT_LIST_DIR}/podio-vis DESTINATION ${CMAKE_INSTALL
 if(ENABLE_RNTUPLE)
  install(PROGRAMS ${CMAKE_CURRENT_LIST_DIR}/podio-ttree-to-rntuple DESTINATION ${CMAKE_INSTALL_BINDIR})
 endif()
+install(PROGRAMS ${CMAKE_CURRENT_LIST_DIR}/podio-merge-files DESTINATION ${CMAKE_INSTALL_BINDIR})
 
 # Add a very basic test of podio-vis
 if(BUILD_TESTING)

tools/podio-merge-files

@@ -0,0 +1,53 @@
+#!/usr/bin/env python3
+"""podio-merge-files tool to merge any number of podio files into one"""
+
+import argparse
+import podio.root_io
+from podio import reading
+
+parser = argparse.ArgumentParser(
+ description="Merge any number of podio files into one, can merge TTree and RNTuple files"
+)
+
+parser.add_argument("--output-file", help="name of the output file", required=True)
+parser.add_argument("files", nargs="+", help="which files to merge")
+parser.add_argument(
+ "--metadata",
+ choices=["none", "all", "first"],
+ default="first",
+ help="metadata to include in the output file, default: "
+ "only the one from the first file, other options: all files, none",
+)
+parser.add_argument(
+ "--metadata-category-name",
+ default="metadata",
+ help="name of the metadata category in the output file, default: metadata",
+)
+args = parser.parse_args()
+
+all_files = set()
+for f in args.files:
+ if f in all_files:
+ raise ValueError(f"File {f} is present more than once in the input list")
+ all_files.add(f)
+
+root_format = reading._determine_root_format(args.files[0])
+if root_format == reading.RootFileFormat.TTREE:
+ reader = podio.root_io.Reader(args.files)
+ writer = podio.root_io.Writer(args.output_file)
+elif root_format == reading.RootFileFormat.RNTUPLE:
+ reader = podio.root_io.RNTupleReader(args.files)
+ writer = podio.root_io.RNTupleWriter(args.output_file)
+else:
+ raise ValueError(f"Input file {args.files[0]} is not a TTree or RNTuple file")
+
+
+for category in reader.categories:
+ if category == args.metadata_category_name and args.metadata == "none":
+ continue
+ if category == args.metadata_category_name and args.metadata == "first":
+ all_frames = [reader.get(category)[0]]
+ else:
+ all_frames = reader.get(category)
+ for frame in all_frames:
+ writer.write_frame(frame, category)