Learned this trick from TSGal just last week:
Use this criterion:
The 2. characters are being interpreted as a word separator and thus *.PDF won't find a "word" that FileMaker interprets as PDF in the case of the last record.
@ Phil: No that doesn't work either if there are any text after the .pdf anytext...
==*.pdf* will thou.
Just tried all thinkable variants and just the one above or just "pdf" got me all, as soon a dot is involved the search engine goes bananas for some reason.
Since we are talking file names here there shouldn't be any text to the right of .PDF or you don't have a valid file name. I did test this with the examples provided and ==*.PDF works for the examples in the original post.
Well, I read it as in a text-field and then... but ur right if that is the only thing in the field.
Well one of the things we do very often when we answer a posted question is guess as to the full context behind the question. Often it's like looking through an old fashioned keyhole of a house's front door and making guesses as to the color scheme used in the bathroom.
Just tried all thinkable versions and just the one above or just "pdf" got me all, as soon a dot is engaged the online look for motor goes apples for some purpose.
To see why, do this:
Set up a table with just the three sample values shown here entered into a text field of three different records.
right click the field and select "insert from index". Then select the "show individual words" option. You'll see that you have Treeroot2 and PDF as two words, not one. This is due to the combination of a numerical digit followed by a period. That period is interpreted as a decimal point and thus the letters after become part of the next word so you get two words in the index where we would otherwise expect to see 1.
Yes, agree, noticed that when I was writing a extractor-script to get some text from a web-site... where currency was some of the "text" retrieved... dots or commas had to be interpreted for... and surely there is other chars that behave the same I guess.
Thanks, everyone. I'll try to widen the keyhole a bit...
It's a text field, but there shouldn't (one of my favorite words) be any text after .pdf, as this IS a catalog of image names. So Phil, your answer does work--tried it. (With the asterisk at the end, it also found a dummy record I created with the name 8397 TreeRoot2.pdfword as CekariYH noted.)
I also tried the test to insert from index and I do indeed see TreeRoot2 and pdf as two words, while TreeRoot.pdf is listed as one word.
So... even in a TEXT field, a number followed by a period causes the period to be treated like a decimal point. And the decimal point is considered a "word" separator.
I still don't quite get the the "word separator" logic... If I search for *.pdf*, I just expect to find any record where .pdf occurs in the field...
Guess I have to rethink my assumption that if I can't find it with asterisks on either side, it ain't there...
Thanks again, everyone.
I agree that it's odd, but it's what we have to work with.