avoid merge if num_files_to_merge is 1

jermp · May 25, 2022 · a5f26c1 · a5f26c1
1 parent 90cc9ff
commit a5f26c1
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 6 deletions.
diff --git a/include/builder/build_index.cpp b/include/builder/build_index.cpp
@@ -72,16 +72,17 @@ struct bucket_pairs {
     }
 
     std::string get_bucket_pairs_filename() const {
+        if (m_num_files_to_merge == 1) return get_tmp_output_filename(0);
         std::stringstream filename;
         filename << m_tmp_dirname << "/sshash.tmp.run_" << m_run_identifier << ".bucket_pairs.bin";
         return filename.str();
     }
 
     void merge() {
-        if (m_num_files_to_merge == 0) return;
-
-        assert(m_num_files_to_merge > 0);
         std::cout << "files to merge = " << m_num_files_to_merge << std::endl;
+        if (m_num_files_to_merge <= 1) return;
+
+        assert(m_num_files_to_merge > 1);
 
         struct iterator_type {
             iterator_type(bucket_pair const* b, bucket_pair const* e) : begin(b), end(e) {}
@@ -164,7 +165,7 @@ struct bucket_pairs {
     std::string m_tmp_dirname;
     std::vector<bucket_pair> m_buffer;
 
-    std::string get_tmp_output_filename(uint64_t id) {
+    std::string get_tmp_output_filename(uint64_t id) const {
         std::stringstream filename;
         filename << m_tmp_dirname << "/sshash.tmp.run_" << m_run_identifier << ".bucket_pairs."
                  << id << ".bin";

diff --git a/include/builder/util.hpp b/include/builder/util.hpp
@@ -208,10 +208,10 @@ struct minimizers_tuples {
     }
 
     void merge() {
+        std::cout << "files to merge = " << m_num_files_to_merge << std::endl;
         if (m_num_files_to_merge == 0) return;
 
         assert(m_num_files_to_merge > 0);
-        std::cout << "files to merge = " << m_num_files_to_merge << std::endl;
 
         struct iterator_type {
             iterator_type(minimizer_tuple const* b, minimizer_tuple const* e) : begin(b), end(e) {}
@@ -305,7 +305,7 @@ struct minimizers_tuples {
     std::string m_tmp_dirname;
     std::vector<minimizer_tuple> m_buffer;
 
-    std::string get_tmp_output_filename(uint64_t id) {
+    std::string get_tmp_output_filename(uint64_t id) const {
         std::stringstream filename;
         filename << m_tmp_dirname << "/sshash.tmp.run_" << m_run_identifier << ".minimizers." << id
                  << ".bin";