IBM DB2 support #25
IBM DB2 support #25
Conversation
Unfortunately I didn't find any other solution for this...
| @@ -57,11 +63,11 @@ export default class Tokenizer { | |||
| // 5. national character quoted string using N'' or N\' to escape | |||
| createStringPattern(stringTypes) { | |||
| const patterns = { | |||
| "``": "((`[^`]*($|`))+)", | |||
There was a problem hiding this comment.
Reordered them to match documentation order.
src/languages/Db2Formatter.js
Outdated
| ]; | ||
|
|
||
| const reservedToplevelWords = [ | ||
| "ADD", "AFTER", "ALTER COLUMN", "ALTER TABLE", "DELETE FROM", "EXCEPT", "FETCH FIRST", "FROM", "GROUP BY", "GO", "HAVING", |
There was a problem hiding this comment.
Only thing that was added is "FETCH FIRST"
There was a problem hiding this comment.
So to this group has "FETCH FIRST" compared to StandardSqlFormatter.
Might be worth to document it. Another thing: would be nice if the lines aligned with StandardSqlFormatter, so I could see differences when I just run diff for these files.
Of course a best diff would be one word per line. Perhaps it's actually worth doing that, although it creates pretty long files, it should be easier to maintain. Like currently I see that there are lots of differences in the reservedWords list, but it's really hard to tell what are the exact differences.
There was a problem hiding this comment.
I tried making toplevel and newline words easier to compare because these are the most important ones.
I think one word per line is overkill. If you really want to compare differenct dialects config then you can use online tools or write a quick script with Lodash (that's what I did).
src/core/Tokenizer.js
Outdated
| */ | ||
| constructor(cfg) { | ||
| this.WORD_REGEX = /^(\w+)/; | ||
| this.WORD_REGEX = /^([\w|#|@]+)/; |
There was a problem hiding this comment.
I tried playing around with word and operators tokenizing order (match operators before words) but that was no-go. Currently I went with the easiest solution that shouldn't break anything...
There was a problem hiding this comment.
We should also make this WORD_REGEX configurable. We could have a config option like specialWordChars: ["@", "#"].
Other SQL dialects don't allow # and @ inside identifiers, which currently results in a scenario where formatting the following query with StandardSqlFormatter:
SELECT a#comment, hereresults in:
SELECT
a#comment,
herewhile correct formatting would be:
SELECT
a #comment, here
| " -- This is a comment\n" + | ||
| " MyTable;\n" | ||
| ); | ||
| }); |
There was a problem hiding this comment.
We should also test the formatting of @ and # inside identifiers. This test partially covers the latter, but it would be better to have a separate test with clear description, explaining the @ and # are part of identifiers (e.g. not treated as operators).
src/core/Tokenizer.js
Outdated
| @@ -49,6 +50,10 @@ export default class Tokenizer { | |||
| return new RegExp(`^(${reservedWordsPattern})\\b`, "i"); | |||
| } | |||
|
|
|||
| createWordRegex(specialChars = []) { | |||
| return new RegExp(`^([\\w|${specialChars.join("|")}]+)`); | |||
There was a problem hiding this comment.
This will result in regex like: ^([\w|@|#]+)
As the | characters are part of character class [ ... ] this will match stuff like foo|bar as a single word, as | has no special meaning inside square brackets.
test/Db2FormatterTest.js
Outdated
| @@ -38,7 +38,8 @@ describe("Db2Formatter", function() { | |||
|
|
|||
| it("recognizes @ and # as identifiers", function() { | |||
There was a problem hiding this comment.
It would be more correct to say: "as part of identifiers"
This got added in the original PR for DB2 support: #25 Don't know how Uku interpreted the DB2 manual, but I for one don't see these characters being allowed in identifiers. Unfortunately I didn't check the manual myself when I originally accepted this PR.
I used ftp://ftp.software.ibm.com/ps/products/db2/info/vr105/pdf/en_US/DB2SQLRefVol1-db2s1e1051.pdf for learning and setting up DB2 configuration.