prestodb/presto · error · PrestoException

DRUID_SEGMENT_LOAD_ERROR

DRUID_SEGMENT_LOAD_ERROR

Error message

Zip doesn't contain file: %s

What it means

Thrown when readFile(name) is asked for a file that does not exist inside the segment zip archive. ZipIndexFileSource reads Druid segments stored as zips and looks up entries in the central directory (zipData); a missing entry means the requested file isn't in the archive.

Source

Thrown at presto-druid/src/main/java/com/facebook/presto/druid/segment/ZipIndexFileSource.java:67

    private ZipFileData zipData;

    public ZipIndexFileSource(DataInputSource dataInputSource)
    {
        this.dataInputSource = requireNonNull(dataInputSource, "dataInputSource is null");
        zipData = readCentralDirectory();
    }

    /**
     * Reads file inside a zip archive
     */
    @Override
    public byte[] readFile(String fileName)
            throws IOException
    {
        ZipFileEntry entry = zipData.getEntry(fileName);
        if (entry == null) {
            throw new PrestoException(DRUID_SEGMENT_LOAD_ERROR, format("Zip doesn't contain file: %s", fileName));
        }
        byte[] fileData = new byte[(int) entry.getSize()];
        readFully(entry, 0, fileData, 0, fileData.length);
        return fileData;
    }

    @Override
    public final void readFile(String fileName, long position, byte[] buffer)
            throws IOException
    {
        ZipFileEntry entry = zipData.getEntry(fileName);
        if (entry == null) {
            throw new PrestoException(DRUID_SEGMENT_LOAD_ERROR, format("Zip doesn't contain file: %s", fileName));
        }
        readFully(entry, position, buffer, 0, buffer.length);
    }

    private void readFully(ZipFileEntry entry, long position, byte[] buffer, int bufferOffset, int bufferLength)

View on GitHub (pinned to 55bb57d202)

Solutions

1. Verify the requested file/column exists (list zip entries: unzip -l segment.zip)
2. Refresh the segment cache or re-download the segment zip
3. Align connector version with the Druid segment format version
4. Fix the column/table mapping to match files actually in the zip

Example fix

// before
byte[] meta = indexFileSource.readFile("__metadata");
// after
if (zipHasEntry(indexFileSource, "__metadata")) {
    byte[] meta = indexFileSource.readFile("__metadata");
}

// caller-side pre-check
ZipFile zip = new ZipFile(segmentZipPath);
boolean present = zip.stream().anyMatch(e -> e.getName().equals("__metadata"));
if (!present) { throw new PrestoException(NOT_FOUND, "Entry missing from segment zip"); }

function hasEntry(entries, name) { return entries != null && entries.some(e => e.name === name); }

try {
    byte[] data = indexFileSource.readFile(fileName);
} catch (PrestoException e) {
    if (e.getErrorCode().getName().equals("DRUID_SEGMENT_LOAD_ERROR") && e.getMessage().contains("doesn't contain file")) {
        // treat as missing column / stale segment: re-fetch
    } else { throw e; }
}

When it happens

Trigger: readFile(fileName) finds no ZipFileEntry for fileName — requesting meta.json/__metadata/a column file that the zip lacks.

Common situations: Querying a column absent from the segment zip, stale segment cache pointing to an older zip layout, wrong segment file supplied, Druid version differences in internal file names.

Related errors

• DRUID_SEGMENT_LOAD_ERROR
• DRUID_SEGMENT_LOAD_ERROR
• DRUID_SEGMENT_LOAD_ERROR
• DRUID_SEGMENT_LOAD_ERROR
• DRUID_SEGMENT_LOAD_ERROR

AI-assisted analysis of prestodb/presto@55bb57d202 (2026-09-04). Data as JSON: /api/errors/2fdadc046591fdb2. Report an issue: GitHub.