Question

I’m having some trouble remembering my database queries – it’s been a while 🙁

Let’s say I have some tables:

WORDS {ID, WORD}
FILES {ID, URL}
FILES_WORDS {WORDID, FILEID, COUNT}

WORDS and FILES both have PK’s named ID and a string for their word/path respectively. FILES_WORDS denormaizes the many-to-many relationship and holds the count of times each word appeared in each file.

I know I can select the information for a given file like this:

SELECT FILES.URL, WORDS.WORD, FILES_WORDS.COUNT FROM WORDS, FILES, FILES_WORDS
    WHERE FILES_WORDS.FILEID = FILES.ID AND FILES_WORDS.WORDID = WORDS.ID AND
    IFILE.ID = 1;

I’m sure there’s a cleaner syntax using joins and things like that though and I’m having trouble googling it or remembering it. Can anyone help speed/clean this up?

Answer 1

I think you are looking for something like this:

SELECT F.URL, W.WORD, FW.COUNT 
FROM WORDS W
  INNER JOIN FILES_WORDS FW
    ON FW.WORDID = W.ID
  INNER JOIN FILES F
    ON FW.FILEID = F.ID
WHERE FW.FILEID = 1;

But it is not necessarily faster. Depends on volume, indices, etc.

I used aliases (the one letter nickname for the table) because it makes it easier to follow, and I don’t think you need both joins, but that also depends on what you are trying to do.