XREAD and Blocking Reads

Following a stream like tail -f

XRANGE reads history; XREAD is designed for following one or more streams. You pass, for each stream, the last ID you have seen, and Redis returns entries with greater IDs. With BLOCK milliseconds, the call waits for new entries if none are available, like tail -f; BLOCK 0 waits forever. The special ID $ means "only entries added after this call", useful for the very first read when you do not care about history. After each batch, remember the last ID returned and pass it in the next call, so you never miss or repeat entries. COUNT limits the batch size. XREAD is for independent readers: every reader sees every entry and tracks its own position. When several workers must share the work, each entry going to only one of them, use consumer groups instead (next section).

A simple stream follower

The reader keeps its own last ID so it resumes correctly after each batch.

last_id = "$"            # start with new entries only (use "0" to read from the beginning)

while True:
    batches = r.xread({"orders": last_id}, count=100, block=5000)   # wait up to 5 s
    for stream_name, entries in batches:
        for entry_id, fields in entries:
            print(entry_id, fields["orderId"], fields["status"])
            last_id = entry_id      # remember position
    # persist last_id somewhere if the reader must resume after a restart

Bookmarking a newspaper archive

XREAD is reading the archive from your bookmark onwards. Each reader has their own bookmark, and reading does not tear pages out, so others can read the same pages.

Quick check: What does the special ID `$` mean in `XREAD ... STREAMS orders $`?

• Read from the very first entry
• Return only entries added after the call starts
• Read the last entry again
• Delete the stream