Method removeImpl

src/Storages/MergeTree/IMergeTreeDataPart.cpp:1499–1647

}

void IMergeTreeDataPart::removeImpl(bool keep_shared_data) const
{
    /// load checksums before move any part files
    ChecksumsPtr checksums;
    try
    {
        checksums = getChecksums();

        /// load checksums for projections parts before removing parent part.
        for (const auto & [ _, projection_part] : projection_parts)
            projection_part->getChecksums();
    }
    catch(const Exception & e)
    {
        if (e.code() == ErrorCodes::NO_FILE_IN_DATA_PART)
        {
            LOG_WARNING(storage.log, "No checksum.txt when getting part's checksums, most likely the part is generated by an aborted task.");
        }
        else
            throw;
    }

    // remove deleted files in checksums to avoid removing the deleted columns/projections
    if (checksums)
    {
        for (auto it = checksums->files.begin(); it != checksums->files.end();)
        {
            const auto & file = it->second;
            if (file.is_deleted)
                it = checksums->files.erase(it);
            else
                ++it;
        }
    }

    /** Atomic directory removal:
      * - rename directory to temporary name;
      * - remove it recursive.
      *
      * For temporary name we use "delete_tmp_" prefix.
      *
      * NOTE: We cannot use "tmp_delete_" prefix, because there is a second thread,
      *  that calls "clearOldTemporaryDirectories" and removes all directories, that begin with "tmp_" and are old enough.
      * But when we removing data part, it can be old enough. And rename doesn't change mtime.
      * And a race condition can happen that will lead to "File not found" error here.
      */

    fs::path from = fs::path(storage.getRelativeDataPath(location)) / relative_path;
    fs::path to = fs::path(storage.getRelativeDataPath(location)) / ("delete_tmp_" + name);
    // TODO directory delete_tmp_<name> is never removed if server crashes before returning from this function

    auto disk = volume->getDisk();
    if (disk->exists(to))
    {
        LOG_WARNING(storage.log, "Directory {} (to which part must be renamed before removing) already exists. Most likely this is due to unclean restart. Removing it.", fullPath(disk, to));
        try
        {
            disk->removeSharedRecursive(fs::path(to) / "", keep_shared_data);