Initial commit

notes/database-prog/sql-scripts/2024-01-24.sql

@@ -0,0 +1,105 @@
+-- T-SQL
+-- Keywords - UPPERCASE, Tables/Columns/etc - PascalCase, everything else - doesn't matter
+-- Comments with double-hyphen
+USE AP
+GO
+
+SELECT VendorID, VendorName, LEFT(VendorName, 10) VendorShortName,
+    VendorAddress1, VendorAddress2, VendorCity
+FROM Vendors
+
+-- batch run everything before it
+-- GO 2
+
+SELECT *
+FROM Terms
+
+SELECT InvoiceID, InvoiceNumber, InvoiceTotal, PaymentTotal, CreditTotal,
+    CASE PaymentTotal
+		WHEN InvoiceTotal THEN 1
+		ELSE 0
+	END IsPaidFull
+FROM Invoices
+
+-- includes credit
+SELECT InvoiceID, InvoiceNumber, InvoiceTotal, PaymentTotal, CreditTotal, InvoiceTotal - CreditTotal PaidTotal,
+    CASE PaymentTotal
+		WHEN InvoiceTotal - CreditTotal THEN 1
+		ELSE 0
+	END IsPaidFull
+FROM Invoices
+
+-- better
+SELECT TOP 500
+InvoiceID, InvoiceNumber, InvoiceTotal, PaymentTotal, CreditTotal, InvoiceTotal - CreditTotal PaidTotal,
+    CASE WHEN PaymentTotal = InvoiceTotal - CreditTotal THEN 1
+		ELSE 0
+	END IsPaidFull
+FROM Invoices
+ORDER BY VendorID, InvoiceDueDate
+
+SELECT TOP 50 InvoiceID, VendorID FROM Invoices
+ORDER BY VendorID
+
+-- 2024-01-29
+
+-- paging
+-- skip first offset rows, return next fetch rows
+SELECT InvoiceID, InvoiceNumber, InvoiceDueDate
+FROM Invoices
+ORDER BY InvoiceDueDate
+OFFSET 110 ROWS
+FETCH NEXT 10 ROWS ONLY
+
+-- Filtering
+-- just return what you need, instead of filtering in the application
+SELECT InvoiceID, VendorID, InvoiceDueDate, InvoiceTotal
+FROM Invoices
+-- Note that equality is a single equals sign
+WHERE InvoiceTotal >= 1000 OR VendorID = 110
+ORDER BY VendorID, InvoiceDueDate
+
+SELECT VendorID, VendorName, VendorCity, VendorState
+FROM Vendors
+-- Note strings use apostrephes, not quotation marks
+-- Keep in mind *there is no short-circuit evaluation*, so this will run every single thing and it's messy and slow
+--WHERE VendorState = 'CA' OR VendorState = 'AZ' OR VendorState = 'OH' OR VendorState = 'NY'
+-- Instead, use IN
+WHERE VendorState IN ('CA', 'AZ', 'OH', 'NY')
+ORDER BY VendorName
+
+SELECT InvoiceID, InvoiceNumber, InvoiceTotal
+FROM Invoices
+--WHERE InvoiceTotal >= 100 AND InvoiceTotal <= 500
+--less typing and more optimized
+--keep in mind it's an inclusive between, not exclusive
+WHERE InvoiceTotal BETWEEN 100 AND 500
+ORDER BY InvoiceDueDate
+
+-- Haven't paid
+-- if either VendorAddress1 or VendorAddress2 are `NULL`, then the concatenation will just be `NULL`
+-- so this filters out all the nulls
+-- (note that the VendorAddress1 one isn't actually needed because the dataset doesn't contain any that has VendorAddress1 as `NULL` but VendorAddress2 defined)
+SELECT VendorID, VendorName, VendorAddress1, VendorAddress2, VendorAddress1 + ' ' + VendorAddress2 FROM Vendors
+WHERE VendorAddress1 IS NOT NULL AND VendorAddress2 IS NOT NULL
+-- or ISNULL to just return a string instead of NULL
+SELECT VendorID, VendorName, VendorAddress1, VendorAddress2, ISNULL(VendorAddress1, '') + ' ' + ISNULL(VendorAddress2, '') FROM Vendors
+-- or COALESCE
+SELECT VendorID, VendorName, VendorAddress1, VendorAddress2, COALESCE(VendorAddress1, '') + ' ' + COALESCE(VendorAddress2, '') FROM Vendors
+
+-- All vendors with PO Boxes
+SELECT VendorID, VendorName, VendorAddress1, VendorAddress2 FROM Vendors
+-- '%PO BOX%' is equivalent to '.*PO BOX.*' in regex
+WHERE VendorAddress1 LIKE '%PO BOX%' OR VendorAddress2 LIKE '%PO BOX%'
+
+-- Get current date
+--SELECT GETDATE(), GET UTCDATE()
+-- Get all invoices for January
+SELECT InvoiceId, InvoiceNumber, InvoiceDueDate FROM Invoices
+WHERE MONTH(InvoiceDueDate) = 1
+--WHERE DAY(InvoiceDueDate) IN (30, 31)
+-- Get all invoices for 2019
+--WHERE YEAR(InvoiceDeDate) = 2019
+-- More efficient version, also more flexible, but longer
+--WHERE InvoiceDueDate BETWEEN '1/1/2019' AND '12/31/2019'
+

notes/database-prog/sql-scripts/2024-01-31.sql

@@ -0,0 +1,19 @@
+-- #### Start of class stuff ####
+
+-- Get the InvoiceId, InvoiceTotal, and InvoiceDueDate for all invoices for vendor 110
+SELECT InvoiceId, InvoiceTotal, InvoiceDueDate
+FROM Invoices
+WHERE VendorID = 110
+
+-- Get the latest 10 invoices for vendor 123: InvoiceId, InvoiceDueDate, InvoiceTotal
+SELECT TOP 10 InvoiceId, InvoiceTotal, InvoiceDueDate
+FROM Invoices
+WHERE VendorID = 123
+ORDER BY InvoiceDueDate DESC
+
+-- Get the InvoiceId, VendorId, InvoiceDueDate, and InvoiceTotal for all invoices due between 1/1/2020 and 1/31/2020
+SELECT InvoiceId, VendorId, InvoiceDueDate, InvoiceTotal
+FROM Invoices
+WHERE InvoiceDueDate BETWEEN '1/1/2020' AND '1/31/2020'
+
+-- ######################################################################## --

notes/database-prog/syntax.md

@@ -0,0 +1,44 @@
+# SQL Syntax
+
+## Keywords
+
+`|`: Use for designating different things that can be used as arguments for the keywords
+`( )`: Optional keywords
+`[ ]`: Used for values, like integers, or column or table names.
+
+| Keyword | Explanation | Example |
+| ------- | ----------- | ------- |
+| `USE` | Specify the database to be used | `USE WideWorldImporters` |
+| `SELECT {ColumnID \| *}`  | selects the column to return (or can use wildcard `*`) | `SELECT ProductID FROM Products` |
+| `FROM` | specify the thing to be selected from |
+| `GO {count}` | Runs everything before it `count` times minus 1 | `GO 5` (runs everything before it 4 more times) |
+| `TOP {n} (PERCENT - optional)` | Returns the top *n* (or *n*%) results - must be an integer . Order is undefined, but in most implementations is listed in the order they were inserted - NOT guaranteed behavior, just how most implementations use it. | `SELECT TOP 10 PERCENT VendorName  FROM Vendors` |
+| `ORDER BY {column1, column2 (1 required, can use infinitely many)}` | Orders by some column (alphanumerically sorted, letters first); `TOP` is useless without this | `SELECT TOP 50 InvoiceID, VendorID FROM Invoices ORDER BY VendorID` |
+| `DISTINCT` | Goes before the column names (and before TOP in MS SQL server, at least); removes rows with duplicate values | `SELECT DISTINCT VendorName FROM Invoices` |
+
+<!-- Needs to be redone into more detailed sub-headings -->
+
+## `CASE`/`WHEN`/`THEN`/`ELSE`/`END`
+
+## Other syntax
+
+- Brackets `[ ]`: Delineates that something is a table, usually not needed but allows for stuff like using keywords as table names.
+
+## Other notes
+
+- `SELECT * FROM Table`: Shouldn't be used normally, only in learning or if you're not sure what you need yet.
+  - Bad because: it's requests every column of data in the table, and that's rarely needed, wasting time and resources.
+  - Most database systems will just skip all qury optimizations if you do this, wasting even more time and resources.
+  - Column ordering is undefined - the columns could be returned in any random order.
+- Should usually do one keyword per line, except for simple stuff like `SELECT ColumnID FROM Table`
+- Column aliases: Useful for:
+  - Multi-table connections and for "renaming" columns
+  - Making it more convenient to use (e.g. `Person` -> `Name` as an alias of `Person` -> `PersonName`)
+  - To name an unnamed column (e.g. `SELECT LEFT(VendorName, 10) VendorShortName FROM Vendors`)
+
+## No
+
+Don't:
+
+- Put spaces in a table's name (why would you???)
+- Use `SELECT * FROM Table`