Menu

#68 Word Query listing without frequencies

open
Client (37)
5
2007-02-12
2007-02-12
No

When I save a listing from the Word Query box, I'm getting very odd results (working with 1 million word corpora, but only downloading freq < 99).

For instance:

<wordList lemma-scheme="null">
<pattern>.*</pattern>
<entry freq="58532"><lemma><word>.</word></lemma></entry>
<entry freq="55251"><lemma><word>,</word></lemma></entry>
<entry freq="37796"><lemma><word>the</word></lemma></entry>
<entry freq="31751"><lemma><word>i</word></lemma></entry>
<entry freq="26141"><lemma><word>you</word></lemma></entry>
<entry freq="25661"><lemma><word>it</word></lemma></entry>
<entry freq="24770"><lemma><word>and</word></lemma></entry>
<entry freq="20949"><lemma><word>that</word></lemma></entry>
<entry freq="20780"><lemma><word>to</word></lemma></entry>
<entry freq=""><lemma><word>'s</word></lemma></entry>
<entry freq=""><lemma><word>a</word></lemma></entry>
<entry freq=""><lemma><word>?</word></lemma></entry>
<entry freq=""><lemma><word>of</word></lemma></entry>
<entry freq=""><lemma><word>n't</word></lemma></entry>
<entry freq=""><lemma><word>in</word></lemma></entry>
<entry freq=""><lemma><word>we</word></lemma></entry>
<entry freq=""><lemma><word>er</word></lemma></entry>
<entry freq=""><lemma><word>do</word></lemma></entry>
<entry freq=""><lemma><word>is</word></lemma></entry>
<entry freq=""><lemma><word>they</word></lemma></entry>
<entry freq=""><lemma><word>!</word></lemma></entry>
<entry freq=""><lemma><word>yeah</word></lemma></entry>
<entry freq=""><lemma><word>on</word></lemma></entry>
<entry freq=""><lemma><word>was</word></lemma></entry>
<entry freq=""><lemma><word>have</word></lemma></entry>
<entry freq=""><lemma><word>what</word></lemma></entry>
<entry freq=""><lemma><word>he</word></lemma></entry>
<entry freq=""><lemma><word>there</word></lemma></entry>

....

IE all the frequency counts for items lower down the list are missing.

It doesn't happen if, prior to saving the listing, I scroll slowly down through the list. Presumably doing this is forcing the freqs to load. But surely saving a listing should force them all to load anyway.

Andrew.

Discussion

  • Lou Burnard

    Lou Burnard - 2007-02-26

    Logged In: YES
    user_id=1021146
    Originator: NO

    Saving a listing does download all the hits, but it takes a while. If you get something like this behaviour *after* the little red light goes out/hourglass disappears then something is wrong. Your example above shows lots of hits with frequency > 99, which looks as if the download is seriously incomplete. Don't forget that the frequency isn't stored in the index, so the client still has to download them all before it can decide whether they should go into the listing.

     
  • Andrew Hardie

    Andrew Hardie - 2007-02-27

    Logged In: YES
    user_id=1460495
    Originator: YES

    Sorry, I meant I'd restricted it to freq > 99, althoguh I've tested it with the full freq list now and I'm getting the same problem even if I don't put any restriction on.

    I *am* getting this behaviour after the little red light disappears, in any case (I don't actually get the red light for this, just a half-second hour glass). Not only am I getting it after the hour glass disappears, I'm getting it after I let the hour glass disappear, go away for two or three minutes to read a webpage, then come back! If I leave it for ten minutes or so then it does eventuially download them all.

    So it looks as if something is indeed wrong. Basically what I'm arguing is that it shouldn't allow me to save a listing until all the info it needs for that listing is ready to be written.