Author: JS

Grep

I use grep quite a bit in BBEdit for simple replacement but nothing especially complicated. Recently I have been cleaning up my PHP error log after installing PHP 8.1 after years of using PHP 5. Most of the changes were easy, just initializing variables before use. Some were more complicated like getting rid of braces in array calls. And some took a while to figure out, like accessing session variables. I also had some beta code that I didn’t want to address yet. So the grep commands I settled on were variants of this.


grep -ihrw '/var/log/php_error.log' -e 'locutour' | grep -v '$_SESSION' | grep -v 'array'

-i is case insensitive
-h supresses the file name, which I already know
-r is a recursive flag, which I don’t need in this particular instance
-w matches the whole word

-e is followed by the pattern or patterns, in this case looking for just one website. You can include multiple patterns with multiple instances of -e e.g. -e 'locutour' -e 'wellgolly'

I didn’t feel like dealing with session and array warnings on the first pass, so I piped the results to a grep with the -v flag to remove those warnings.

Because it is a pipe, you can redirect it to a file instead of the screen and I did that when looking for words in a file of Nero Wolfe books when compiling a database of my favorite Nero Wolfe words. In this case I was looking for all of the adverbs that Archie used when describing how someone spoke. I piped it to a file and used BBEdit to clean it up so that I only had one instance of each.


grep -ihrw  the-complete-NW-modified.txt  -e 'said.*ly\.' > saidly.txt

SQL for managing wordlists

A long time ago I started a notebook where I would record passages from books. Most of them are now in the quotes database. I also started collecting interesting words. The first word whose definition I recorded was fatalism: belief that everything is controlled by fat. The first serious words were bellicose: warlike and ameliorate: to make or desire to be better, improve. There are now 1896 interesting words in the database, along with examples of usage.

As part of an app that we were making we assigned a difficulty level to words. There are over 16,000 words that have a difficulty level assigned to them. The code to assign levels to is straightforward since the tables are in the same database.


UPDATE favorite_words, words_by_level 
SET favorite_words.level = words_by_level.level
WHERE favorite_words.word = words_by_level.word

Even though only one table is being updated, you need to mention both in the UPDATE section of the code.

I have another project where I am collecting words and quotes from Nero Wolfe stories. In that case I want to make sure that every word in that table has a definition in the favorite words table. I’m running the code in the Wolfe database and the favorite word table is in a different database so I need to make sure that I prefix the tables in the other database with the database name.


SELECT NWquote.word FROM NWquote 
LEFT JOIN words.favorite_words 
ON words.favorite_words.word = NWquote.word 
WHERE words.favorite_words.word IS NULL

I looked for words in my Wolfe database that did not have a level in my favorite words database and manually updated the level in the favorite words database.


SELECT DISTINCT NWquote.word FROM NWquote
LEFT JOIN words.favorite_words
ON words.favorite_words.word = NWquote.word
WHERE words.favorite_words.level = 0
ORDER BY NWquote.word

Surprisingly, around two hundred of my favorite words still did not have a difficulty level assigned so after manually checking them I assigned the rest to level 6.


UPDATE words.favorite_words, NWdatabase.NWquote
SET words.favorite_words.level = 6 
WHERE words.favorite_words.word = NWdatabase.NWquote.word 
AND words.favorite_words.level = 0

Sorting columnar data in BBEdit

I had a list of books in a MariaDB table that I wanted to redo by date. There may be a way to export selected fields and sorted data from MariDB using PHPMyAdmin, but I don’t know the trick.

But it is possible, in some cases, to sort data by columns in BBEdit if the data cooperates. In this case it did.


(1, 'Fer-de-Lance', NULL, 1934, '553385453', 
(2, 'The League of Frightened Men', NULL, 1935, '553385453', 
(3, 'The Rubber Band', NULL, 1936, '553763091', 
(9, 'The Silent Speaker', NULL, 1946, '553234978', 
(10, 'Too Many Women', NULL, 1947, '553250663', 
...
(34, 'Black Orchids', NULL, 1942, '0553257196',
(35, 'Not Quite Dead Enough', NULL, 1944, '0553261096',

All I need to do was figure out a grep pattern that would match just the date field.
The first thing I did was search for the field I wanted to sort by. It took some experimenting but the code used to find the field I was looking for is:

NULL, \d{4}

where \d means look for digits and {4} tells how many.

Sort search example
Putting that code into the “Sort using pattern” sort field results in lines sorted by date. Alternatively,


\d{4},

also works.

Then its a simple matter of stripping off the ID field and renumbering.

That word, it does not mean what you think it means.

I’ve noticed a lot of spam and tweets lately that are grammatically correct, but do not quite convey the message the writer thinks they do. This post is a collection of my favorites—it will grow over time.

I work as a professional Web Designer / Developer in a reputed firm.

reputed
1. Accorded a reputation.
2. Supposed or assumed to be true.
3. Generally believed to exist, but that is not the case.

Pretty sure they meant reputable, but who knows.