Change Log

2.0 TODO

Use unfetch, sockette and Websocket Node for new client/server code.
Add SDL schema support. Link.

[2.3.6] 08-12-2019

Fixed issue 161, join queries weren't working as expected with AS queries.
Fixed issue 162, compound select with join/graph queries is now working as expected.
Fixed issue 162, offset/limit are working as expected on fast query paths now.

[2.3.5] 07-23-2019

BREAKING CHANGE Removed WIP transaction API, to be implemented in history plugin.
Updated dependencies.
Fixed issue 155, added create database documentation.
Fixed issue 156, events created before a database is setup will now work as expected.
Fixed issue 141, rawImport now uses transactions for bulk queries to increase performance when possible.
Fixed issue 159, added documentation for .union queries.
Added new batch API in the adapters for improved raw import performance.

[2.3.4] 06-24-2019

Updated dependencies.
SnapDB now honors path variable in config.
Added transaction api. Not all adapters support transactions, each one will need to implement it's own transaction methods.
Added throttle to count update to improve performance.
Added serve-link run script for npm link situations.
Added select option to table creation to mutate rows as they are selected.
Added feature to createDatabase types to implement custom types with a function instead of with an object.
Added immutable property to column definitions, allows you to prevent column updates after the initial insert.
Added .updateImmutable() query argument to force immutable column updates.
Added custom function type definitions to take care of.

[2.3.3] 5-20-2019

Fixed issue with rebuilding list type indexes
Fixed issue with empty string values in secondary indexes.
Added console warning for breaking change in previous release.
Made a few performance improvements.

[2.3.2] 5-16-2019

BREAKING CHANGE RocksDB has been replaced by SnapDB as the default NodeJS store for PERM. See the migration guide for steps to restore the old behavior.

Moved window["@nano-sql"].core to window["@nano-sql/core"] for webpack externals compatibility.
Added RocksDB & LevelDB adapters to readme.
Removed all node dependencies.
Resolved issue 138 and previouse other issues with RocksDB switch.
Hopefully resolved issue 139 by removing all node dependencies. Unable to test since Nativescript doesn't work with NPM link...

[2.3.1] 5-10-2019

Removed secondary index queue.
Fixed issue with falsey secondary index values.
Added optional wasm index to RocksDB adapter. Increases offset/limit and index request performance exponentially.
Fixed a few small bugs.
Restored row filter feature.

[2.3.0] 5-09-2019

Fixed issue with inserts.
Removed query queue and implemented row & index locking.
Added fuzzy search plugin.
Working on moving RocksDB to external plugin.

[2.2.9] 5-03-2019

Fixed a few small issues with the new database API.
Added a new total query to quickly get the total number of records in a table.
If a column value is not passed, the default behavior is now to ignore the column instead of defaulting to a value.
Updated dependencies.

[2.2.8] 4-22-2019

Fixed issue with create table query.
config option is now optional.
Added method to list databases.
Fixed issue with nested function calls.
Added FORMAT_NUMBER function.

[2.2.7] 4-19-2019

Fixed an issue with select table code.

[2.2.6] 4-19-2019

Added new multi database API. Events, queries, TTL, indexes and tables are all scoped to each database.
Fixed issue with create table query requiring table argument passed in.
connect method is now createDatabase. Old connect method will be kept around to prevent breaking changes.

[2.2.5] 4-14-2019

BREAKING CHANGE Preset query api has been updated, check the migration docs out.
Added integrity key to releases to prevent the CDN from silently tampering with future releases.
Improved performance of IndexedDB index requests.
Added table id to describe table queries.
Resolved issue #137, rewrote date type behavior to be more intuitive.
Updated dependencies.

[2.2.4] 4-1-2019

Resolved issue #85.

[2.2.3] 3-25-2019

Switched to VuePress for documentation and main website.
Fixed Issue #135, js math 'max' and 'min' functions no longer overwrite the built in aggregate functions.
Fixed Issue #136, empty lines in a CSV file are ignored.
JS Math functions weren't being calculated correctly, that is now fixed.
Added better handling for undefined in secondary indexes.
Issue #132 was still not working reliably, implemented a custom sorting algorithm to make multi column sorting deterministic.

[2.2.2] 3-24-2019

When using .stream the complete and error callbacks are now optional.
Sort By & Order by performance has been improved.
Fixed Issue #132, multiple column order by now works consistently.
Fixed Issue #133, indexes are now explicitly dropped before the table.
Fixed Issue #134, aggregate functions will return an empty result if there are no rows selected.
Updated dependencies.
Improved error handling for upsert queries.

[2.2.1] 3-6-2019

BREAKING CHANGE The default minified build is now ES6 since browser support has grown significantly. There's still an ES5 friendly build generated at nano-sql.min.es5.js in the same folder.
Added support for function based column defaults:

nSQL().query("create table", {
    name: "posts",
    model: {
        "id:uuid": {pk: true},
        "title:string": {},
        "date:int" {default: () => Date.now() },
    }
}).exec()

[2.2.0] 2-17-2019

Made some optimizations to JOIN code, joins are faster now.
Fixed issue #128. Nested joins based on previous join value is now possible.

[2.1.9] 2-12-2019

BREAKING CHANGE The filter "adapterCreateTable" now uses table instead of tableName to keep with the naming convention of other filters.
Fixed issue #126 with .from usage.
Fixed issue #127, official date type is now supported.
Added features requested in issue #125. mode can now be set on a per table basis and copying rows from one table to another is easier.
performance property of events is now per row and doesn't grow as the query progresses.
Added new clone query type to copy tables from one adapter to another.

nSQL("users").query("clone", {
    mode: new SyncStorage() // or any nanoSQL adapter
}).then(() => {
    // clone table done
})

Added new copyTo query argument. Allows a query result to be streamed directly into another table.

nSQL("orders").query("select", ["COUNT(*) AS totalOrders", "AVG(total) AS avgOrder"]).where(["orderDate", ">", Date.now() - (24 * 60 * 60 * 1000)]).copyTo("orderStats" /* table to copy results into */).exec().then...

[2.1.8] 2-8-2019

Fixed bug with LIKE syntax not working similar/identcal to SQL databases. Added tests.
Fixed types with exporting events.
Fixed events not exporting correctly.
Added autocomplete feature to secondary indexes. Works on all existing secondary indexes that are of string type. Usage: nSQL("table").query("select").where(["column", "LIKE", "search%"]).exec().
Added new INCLUDES LIKE where condition. Useful for matching array columns against a like condition.
New INCLUDES LIKE query works for autocomplete on string[] indexes.
Added nativescript adapter.

[2.1.7] 2-7-2019

Adjusted bug with React Native support.
React native adapter added.

[2.1.6] 2-6-2019

Added new secondary index prop: ignore_case.
Added redis index to readme.

[2.1.5] 2-4-2019

Removed memory caching from secondary index class.
Fixed issue with create table not working if you didn't have the table selected with nSQL.
Fixed issue with show tables query returning redudant results.

[2.1.4] 1-31-2018

Adjusted how table ids are assigned.
Merged PR 21 that added change events to row updates.
Added Cordova Adapter.

[2.1.3] 1-30-2019

Added feature detection for WebSQL in safari to fix issue 115.
Added new filter "configTableSystem" to resolve issue 117.
Fixed issue with new query listener feature.
Used getAll api in IndexedDB to dramatically increase performance of select statements in browsers that support the api. Resolves issue 118.

[2.1.2] 1-30-2019

BREAKING CHANGE Removed .streamEvent function, you can now just pass a boolean into the .stream function to enable events: .stream(onRow, complete, error, useEvents).
objectsEqual utility function is now an alias of the fast-deep-equal library, should have better performance for row comparisons and index updates.
Restored observer query feature as .listen() queries.
Adjusted RocksDB adapter so it can accept other leveldown adapters easily.

[2.1.1] 1-24-2018

Added gitter chat link.
Added foreign key feature and its documentation.
Finished new queries feature and its documentation.
Fixed bug with new row data not returning on row updates.
Fixed a few bugs related to memoization not getting invalidated properly.

[2.1.0] 1-22-2018

Fixed typos in readme file.
Adjusted CLI so passed in arguments overwrite file options.
Added new queries option to tables to replace old action/view api.
Added optional parallel argument to loadJS and loadCSV.
Shifted CLI exported interface names from IusersTable to ItableUsers.
Added types object to CLI build.
Changed types property from array of strings to datamodel object.

[2.0.9] 1-21-2018

Fixed ReadMe typo.
Added .distinct() query argument and tests.

[2.0.8] 1-21-2018

Fixed a typo in the readme.
Fixed package.json error.

[2.0.7] 1-21-2018

Fixed issue 111, added tests to prevent the issue from happening again.
Resolved a few other minor issues with group by.
Added CLI and documentation for it.

[2.0.6] 1-18-2018

Adjusted table ids to be not as pointlessly long.
Added new adapters to readme.

[2.0.5] 1-17-2018

Fixed issue with nan function.
Added new adapters to readme.
Added option to RocksDB adapter to disable caching indexes.

[2.0.4] 1-16-2018

Adjusted base SQLite query function to support streams.

[2.0.3] 1-15-2018

Readme tweaks, added more adapters.

[2.0.2] 1-15-2018

Added DynamoDB and SQLite Adapter support.
DS_Store pull request.
A few readme tweaks.

[2.0.1] 1-15-2018

Complete rewrite of the whole library from the ground up. Previous plugins and adapters developed for nanoSQL 1.X will NOT work for 2.X versions.
Changed config, table model, view, and action setup format. View Migration Docs.
Removed all ORM features, tries, history, observables and fuzzy search from core.
Added CROW function as well as geo spatial index feature for fast CROW queries.
You can now upsert nested values. nSQL("Table.column.nested.value").query("upsert", newValue).where([...]).exec()..
You can now return the query event instead of the result by passing true into .exec()...
Added support for .graph queries, they work very similar to the old ORM system but are far more flexible.
Added support for querying tables returned from a promise. nSQL(() => fetch("records.json").then(d => d.json())).query("select")..stuff....exec().... This also lets you nest nanoSQL queries.
Added support for querying promise based tables in union, join and graph queries.
Added support for nested indexes and array indexes.
Added filter system with over 40 hooks to control and modify data/query flow with plugins.
New Stream API allows data to be streamed directly from the database without caching result sets into memory. nSQL().query("select")...stream(onRow, onComplete, onError). Results will be cached regardless if Group By, Order By, and/or aggregate functions are used.
Upsert/Delete queries no longer load their entire result set into memory before performing their action, uses the stream api internally.
Added UNION query support.
limit().offset() queries now do the same as the .range() query modifier use to do when there aren't orderby/groupby arguments.
.range() is also no longer supported.
union, join and graph queries now use indexes when possible/available.
Custom query functions can now conditionally index columns/values. Check out example CROW function: source.
Events can now target nested row values. nSQL("Table.column.nested.value").on("change", () => {}).
Plugins and adapters can declare core/plugin version dependencies which are validated at runtime.
Added new index offset feature for secondary indexes. Primary keys (used for the secondary indexes) often don't support negative values, the offset feature lets you index negative values that are less than the offset value provided in the index config.
Swapped out LevelDB for RocksDB.
Primary key indexes are no longer kept in memory for better performance.
Database adapters now support secondary index specific features.
Query types that are also supported by SQLite are tested against SQLite for consistency.
All javascript Math functions are supported in queries.
Changed OrderBy and GroupBy syntax: nSQL("table").query("select").orderBy(["col1 ASC", "col2 DESC"]).exec()....
OrderBy and GroupBy now support query functions: nSQL("table").query("select").orderBy(["UPPER(col1) DESC"]).exec()..
All functions are supported in SELECT or WHERE arguments. nSQL("table").query("select", ["UPPER(name)"]).exec()... or nSQL("table").query("select").where(["UPPER(name)", "=", "NAME"]).exec()...
Secondary indexes, primary indexes and slow queries can now be mixed in the same WHERE condition and still get good performance. Secondary indexes and primary key queries must come before slow queries to see the speed improvements.
OrderBy & GroupBy now use fast indexed sorting IF the column being sorted is indexed.
You can now cache result sets and paginate the cache at will. Cached results do not get updated and represent a copy of the data at the time it was read.
Events are now triggered per row and not per query. For example, if a query modified 100 rows you would have previously gotten 1 event from the query notifying you of the 100 changed rows. You will now get 100 separate change events. Each database event has the query object that triggered the event in the event data, allowing you to group together events from the same query.
You can now use .into, .on, and .from as a table selector in your queries: nSQL().query("upsert", {data}).into("myTable").exec().then...
.from can be used to alias one table to another name in join, graph, and union queries. nSQL().query("select").from({table: "myTable", as: "posts"}).join(...).exec()....
.from can also be used to load table data from a promise and alias it to a table name: nSQL().query("select").from({table: () => Promise, as: "posts"}).join(...).exec()....
connect event now fires when the database adapter connects, not when nanoSQL is ready. There is a new ready event that fires when queries are ready to go.
You can now drop and add tables to nanoSQL while it's running using queries.
Added new conform rows query that pulls every row from a given table, conforms it to the current data model, then puts the row back.
Added new rebuild indexes query to trigger index rebuilds.
nanoSQLInstance is now just nanoSQL.
window["nano-sql"] is now window["@nano-sql"].core.

[1.8.1] 1-7-2019

Fixed issue with WebSQL attempting to update existing rows in some cases.

[1.8.0] 12-16-2018

Fixed issue #97, observers shouldn't fire after unsubscribes.
Fixed issue #94, SQLite related adapters now work with primary key selects when using numbers.
Fixed issue #85, problem was identical to issue 94.
Fixed "x row(s) modified" bug.

[1.7.9] 11-10-2018

Fixed secondary index truncation.
Fixed issue #89, iOS/WebSQL support fixed.

[1.7.8] 10-10-2018

Fixed CDN Link issue.

[1.7.7] 10-10-2018

Fixed issue [#88](SQL Error: number of parameters does not match argument count) and #87. WebSQL queries were too large and are now batched.
Fixed issue #85. Suboptimized queries where failing.

[1.7.6] 8-10-2018

Fixed issue with CSV export headers.

[1.7.5] 8-6-2018

Added INCLUDES as HAVE alias for queries.
Fixed issue 79. There's now a sanity check for joins to make sure the where condition is formatted correctly.
Fixed issue 72. Secondary indexes with cache enabled now rebuild correctly.
Fixed issue 76. You can now use the prop unique() in the data models to force nanoSQL to prevent multiple rows from having the same value.

[1.7.4] 8-1-2018

CSV import and export now follow official CSV spec.

[1.7.3] 7-23-2018

Merged PR 74.
Fixed stack call error on database read.

[1.7.2] 7-22-2018

Updated .npmignore according to recommendation by @vladimiry on github.
Merged PR 73, improving promise behavior of exec method.
Added changes to PR to prevent breaking changes of query filter.
Restored query cache.

[1.7.1] 7-19-2018

Better error handling.
Resolved issue 70, added a database versioning/migration code.
Resolved issue 69 and also added a test to prevent the bug from popping up again.
Resolved issue 63, added cursor feature to queries.
Resovled issue 64, added TTL feature.

[1.7.0] 7-3-2018

Merged PR from dtiem that increased limit/offset performance by 17%.
Combined NanoSQL into a monorepo as suggested in issue 58.
Fixed issue 59 regarding primary key write errors not propagating.
Fixed Issue 57, IndxedDB versions can now be specified with idbVersion in the config object.

[1.6.9] 6-14-2018

Removed HAVE from sanity check.
Merged PR 54 for LevelDB optional requires.
Removed src folder from .npmignore becuase it's needed by the new source maps.

[1.6.8] 6-13-2018

Added sanity checking for where queries that require an array value.
Issue 49 moved LevelDB to optional dependencies.

[1.6.7] 6-12-2018

Issue 49 actually resolved this time with new LevelDB API.
Resovled Issue 52, added .npmignore file as suggested.
Fixed typedoc build.
Sourcemaps are now included in the build.

[1.6.6] 6-6-2018

Fixed an issue with secondary indexes not writing on empty string value.

[1.6.5] 6-5-2018

Added support for deeply nested where statements, used to group conditions together. Like this:

nSQL("table").query("select").where([
    ["age", "=", 23], "AND" [
        ["favoriteColor", "=", "blue"], "OR", ["favoriteColor", "=", "orange"]
    ]
]).exec()...

[1.6.4] 6-4-2018

Disabled query cache again as it's causing issues with offset/limit. Fixed 50.
Added optional constructor argument to LevelDB adapter for issue 49.

[1.6.3] 5-27-2018

Using cache secondary index performance penalty is now only 10-50% depending on the adapter being used.
Query queue is only used when cache is disabled now.
Restored cache feature with peer mode.
You no longer need to call connect() before performing queries. The queries will now wait until connect() completes to execute.

[1.6.2] 5-24-2018

Fixed 46 with LevelDB not handling auto incriment correctly.
Fixed 45 with IndexedDB not reconnecting if new databases were added to an existing indexed db store.
Progress on 47. Secondary index performance increase substantially, went from 10-30x to 3-5x.
Fixed bug where passing in a non existent table to "describe table" query would lead to an error. Now it returns a valid response.
Fixed bug where passing falsey values into secondary index didn't work, now the secondary index will only fail to index undefined values.

[1.6.1] 5-20-2018

Added observable queries with a small collection of observable modifiers.
Cache feature has been restored, #34 example is now working.
Added multitab syncing for web browser based databases.
Added peer mode where multiple tabs can communicate events about the same database. Only works in the browser and should be used on combination with IndexedDB/WebSQL/LocalStorage. Example Usage: .config({peer: true, mode: "PERM"}).
Finally found the problem with mangling the props, restored that and it's reduced the browser build size by quite a bit.
Added tests.

[1.6.0] 5-19-2018

Fixed an issue with search() queries not working anymore.
Now using better, more reliable queue implementation courtesy of queue.
Updated all dependencies with ncu -u.
Added tests for new search feature.
Improved LIKE query behavior, matches much closer to MySQL behavior now.
Fixed primary key range select.

[1.5.9] 5-18-2018

Fixed secondary index range select issue.
Added secondary index range tests.
Better upsert performance on WebSQL.

[1.5.8] 5-17-2018

Fixed issue 43, secondary indexes, ORM queries and denormalization queries are now ran in a queue to prevent conflicts.
Cleaned up build system a little.

[1.5.7] 5-15-2018

Fixed LIKE queries when column value isn't string.

[1.5.6] 5-14-2018

Fixed LIKE queries when column value is undefined.

[1.5.5] 5-13-2018

Added feature to allow multiple joins in a single query. 41.
Fixed issue 42 with range queries on secondary indexes.

[1.5.4] 5-2-2018

Fixed an issue with the default database index.

[1.5.3] 5-1-2018

Added ns() prop to disable primary key sorting, useful for speeding up inserts on large data sets.
Fuzzy search indexes don't use sorted indexes by default now.
Added stop word index to fuzzy search system tokenizer.
Improved tokenizer to normalize numbers and fractions.
Added crow() query capability, now you can grab rows by their distance from a given GPS point. Usage: nSQL("table").query("select").where(["crow(latitude, longitude)", "<", 3.5]).exec(). Defaults to kilometers for the distance. Rows must have columns of lat and lon to work. If lat and lon are secondary indexes the query uses an optimized read path and goes MUCH faster.
Fixed bug with NULL and NOT NULL.
More work on the new readme.
Added new whereFunction capability used by levenshtein and crow features. Developers will be able to add their own where functions as they want.

[1.5.2] 4-24-2018

Added better error handling for IndexedDB.
Added support to upsert multiple records when using an array: nSQL("table").query("upsert", [{name: "billy"}, {name: "joel"}]).exec();
Fixed issue #37 with the database not connecting when history is enabled.
Fixed issue #36 with ORM queries not working against non array values.
Fixed issue #34, cache was not disabeling when cache:false was passed into the config object.
Cache is temporarily disabled until I can find the root cause of the cache issue described in #34.

[1.5.1] 4-22-2018

LIKE now uses MySQL LIKE syntax with % and _ characters.
Added levenshtein query operator to allow you to get rows based on levenshtien distance. Usage: nSQ("users").query("select").where(["levenshtein(wordToCompare, column)", "<", 4]).exec()..
Moved to a different, much faster (3x), Levenshtein implementation.
Fixed .range() implementation to match .limit().offset() behavior, added test to make sure it stays that way.
Added is connected check on each query to return better error messages.
All errors now show nSQL prefix.
IndexedDB no longer has a "Webworkers" option, this reduces the build size and increases security since we no longer need to use blob workers or eval.
Added a new browser based test for all the browser based adapters.
Updated binary search again, it wasn't working with the new adapter tests.

[1.5.0] 4-18-2018

BREAKING CHANGE (Custom Function API), PLEASE READ THE MIGRATION GUIDE

Raw import now happens incrimentally and has a onProgress callback.
Rebuilding search index and secondary indexes now have onProgress callback.
default was overwriting at incorrect times, fixed bug and added tests.
Functions weren't working on JOIN queries, fixed and added tests.
New faster binary search implimentation.

[1.4.9] 4-17-2018

Fixed search bug with null rows.
Fixed CSV import bug.

[1.4.8] 4-16-2018

Added levenshtien distance to fuzzy search relevance calculations.
Added further optimizations to the new elastic search features.
Added MySQL Adapter and React Native to README.
Restored default column behavior and added test for it.

[1.4.7] 4-15-2018

Fixed issue with LevelDB and search feature.
Fixed issue with search feature and optimized query checking.
Forgot to do return this on .rowFilter.

[1.4.6] 4-15-2018

Fixed CSV Export and Import formating.
Restored .rowFilter feature.
Added fuzzy search capability to elastic search features.

[1.4.5] 4-15-2018

Fixed and restored query cache features.
Added elastic search style feature, where rows can be indexed, words stemmed, and searches happen much faster. Will add testing and documentation later as the feature is found to be stable.

[1.4.4] 4-14-2018

Fixed .where() statement bugs, added a bunch of test conditions to prevent this kind of thing from popping up again.
Fixed issue with _util table.

[1.4.3] 4-9-2018

Moved package.json dependencies into their appropriate sections.
Added a new .extend("flush") command to clear all contents of the database.

[1.4.2] 4-7-2018

Fixed minifaction of IndexedDB web worker in build.
Better performance on range selects.
Added .disconnect() method.
Table scan .where() conditions are much faster now.
Fixed for React Native compability.
Adjusting the props API: "pk" is now "pk()", "ai" is now "ai()", etc..

[1.4.1] 4-5-2018

Added onProgress function to loadCSV and loadJS methods.
Trie objects are populated much faster now on connect.
A few minor performance and code optimizations.
Typedoc now works with newest version of Typescript! Joyously restored Typedocs.
Updated all dependencies.
Cleaned up comments a bunch since they're visible in Typedocs again.

[1.4.0] 3-23-2018

Fixed denormalization toColumn bug.
Fixed aggregate functions returning nothing if there are no result rows.

[1.3.9] 3-20-2018

Added ability to debounce denormalization calculations.
CSVs are handled much better now.
Added optional filter function argument to .loadCSV method, lets you adjust the rows after they've been parsed from the CSV and before they're put into the db.

[1.3.8] 3-19-2018

Added way to access data models for each table.

[1.3.7] 3-18-2018

Added aggregate denormalization features.

[1.3.6] 3-6-2018

Fixed an issue with event triggering.

[1.3.5] 2-17-2018

Fixed an issue with sub optimized queries.
Added optional disconnect method to adapters.

[1.3.4] 2-13-2018

Fixed issue with .extend("clone"...).
Now supports secondary adapters for live backups.

[1.3.3] 2-11-2018

Removed unnecessary argument from database adapter implimentation.
Working on new readme.
Working on denormalization features.
Added wildcard column {key: "*", type: "*"} that disables column sanitization.

[1.3.2] 2-9-2018

Removed the query pool feature for now for better stability.
Fixed issue with WebSQL batch and range reads.

[1.3.1] 2-6-2018

Added Redis to supported database adapters.

[1.3.0] 2-4-2018

Fixed an issue with scoping on loadJS and loadCSV

[1.2.9] 2-3-2018

Added storage adapter passthrough. Can be used with .extend("get_adapter").

[1.2.8] 2-2-2018

Fixed issue with primary key detection.

[1.2.7] 2-2-2018

Fixed an issue with optimized reads.

[1.2.6] 1-29-2018

Started work for adding denormilzed view feature.
Updated fastALL to use Promise.all.

[1.2.5] 1-28-2018

Added methods to access views and actions for a given table.

[1.2.4] 1-25-2018

Fixed an issue with secondary index rebuild process.
Fixed an issue with plugin initialization.
Added primary key/secondary index .where() optimization to search within a result set.
Better optmimized the select statement read path.

[1.2.3] 1-22-2018

Fixed an issue with the primary keyless implementation.

[1.2.2] 1-22-2018

Added Google Cloud Datastore adapter to readme.
Fixed issue with .extend("clone").
We support tables without primary keys again (yay!).
Secondary indexes better support range queries with number columns.
UUID format was incorrect, fixed now.
Updated dependencies.

[1.2.1] 1-19-2018

Added new optional batchRead method to adapaters, handles parallel key requests on the same table.
Fixed secondary index range query issue.
.extend("clone") copies over one table at a time now instead of all at once.
Added batchRead method to WebSQL. Fixed bug with WebSQL string based primary keys.

[1.2.0] 1-14-2018

Fixed secondary index rebuild doing redudant work.
Fixed seconday index read issue.
Added new .extend("clone") that lets you copy all data from the current adapter being used to another one.
Fixed an issue with a LevelDB dependency.

[1.1.9] 1-7-2018

Added safety checks on data model creation.
ReadME updates.
Removed depreciated CHAIN and ALL classes. Use fastCHAIN and fastALL instead.
Added groupBy option to ORM queries.

[1.1.8] 1-5-2018

The npm package marked was in package-lock.json, has since been removed.

[1.1.7] 12-31-2017

Fixed nodejs dependency issue.

[1.1.6] 12-31-2017

Added .onConnected() method. You can pass a callback that will either get called right away if the database is already connected or it'll get called once the database connects.
Added .rawDump() method, kicks out the contents of all database tables and their rows.
Added .rawImport() method, lets you import table contents directly without type checking, indexing or anything else fancy.

[1.1.5] 12-27-2017

If you were using webpack/browserify with nanoSQL then your bundler was smelling a nodejs specific dependency and bloating your bundle by about a hundred kilobytes... for no reason (sorry about that). The lib has been adjusted to resolve this.

[1.1.4] 12-26-2017

Wrote a Vue Mixin.
Added .isConnected property needed by Vue Mixin.

[1.1.3] 12-26-2017

Wrote a React HOC and refactored todo example using it.
window["nano-sql"] is now exposed as an entry point.

[1.1.2] 12-25-2017

Removed minify settings to fix history bug.

[1.1.1] 12-25-2017

Fixed bug with event system.
Added integration test for event system.

[1.1.0] 12-25-2017

Over 2x write and read performance in this update.
Updated promise library to get better setImediate polyfill.
ALL and CHAIN methods now return a real Promise.
ALL and CHAIN methods rewritten to improve performance.
Several more performance improvements.
Simple write test for 100 rows: Before: ~35ms, After: ~15ms

[1.0.5] 12-21-2017

BREAKING CHANGE, PLEASE READ THE MIGRATION GUIDE

This release is focused on performance, 2 - 4x increase in query speed should be seen most of the time.
Promise lib now only polyfills, otherwise the native browser/NodeJS Promise implimentation is used. Benchmarks show a double in query performance from this one change...
Removed secondary index queries on upsert if there are none.
Added conditional to skip event logic if it's not needed.
Simple write test for 100 rows: Before: ~140ms, After: ~35ms

[1.0.4] 12-17-2017

Added table types to action/view arguments. You can now use table names as a type in action/view arguments to cast an argument to a specific table's data model.
Fixed a bug with avFilter function.

[1.0.3] 12-15-2017

Updated all history and ORM queries to keep them from changing the table pointer set by the user.
Added NanoSQL Cordova plugin to readme.
Fixed .manualExec() method and query to work correctly.

[1.0.2] 12-15-2017

Added public getConfig() method that allows you to read the config parameters.
Updated to a hashing function that is 3x faster.

[1.0.1] 12-12-2017

Moved adapter tests to their own git repository.
Added selective cache invalidation. Now the cache will store the primary keys for each select statement, then only invalidate the cache when that primary key is affected in some way.

[1.0.0] 12-10-2017

MANY BREAKING CHANGES, PLEASE READ THE MIGRATION GUIDE This build is intended to stabilize the library, increase performance and make it easy to extend in the future. The API for v1.X.X releases will be very stable moving forward.

Complete rewrite of the database engine, ORM system, and history system.
History can now be database wide (default), table wide or row wide.
Implimemented new plugin system with lifecycle events of every query.
The new built in storage engine uses the new plugin system, allowing you to remove/adjust how queries are handled.
Completely removed the old custom backend system and added a new adapter system.
You can no longer run the built in memory db with persistence.
All database adapters now use a sorted B-Tree index to gaurantee consistency.
You can use instance tables with nSQL queries. ex: nSQL([{name: "Bill"}, {name: "Bob"}]).query("select")....
History is no longer enabled by default.
Added WebSQL support. Safari in iOS/macOS will use WebSQL by default.
Added support for running IndexedDB in a webworker.
Added new INTERSECT and NOT INTERSECT .where() comparators. To check if any array values intersect with any values in an array of the database column.
.where() statements now accept a function, much like filter: nSQL("table").query("select").where(row => row.age > 20).exec()
Tables without a primary key are no longer supported, unless it's an instance table.
Added loadCSV method to automate importing CSVs into the database.
You can now get the version of NanoSQL being used with nSQL().version.
Database versions will now be tracked to make automatic migration changes in the future.
updateORM query has been removed entirely.
ORM updates now happen along side upsert, delete and drop queries.
ORM updates are more consistent and way more performant.
Added a large number of integration tests.

[0.9.3] 10-15-2017

Added changed rows property to database events.

[0.9.2] 10-2-2017

Fixed a bug with the event system running into an error and preventing further errors from being fired.

[0.9.1] 8-26-2017

Added "NOT LIKE" and "!=" comparisons for queries. Not sure how I got by without using these for so long.

[0.9.0] 6-21-2017

Added "NULL" and "NOT NULL" as valid query values; example: ["column", "=", "NULL"].
Changed type casting system to allow null values into the database if no default is set.
.length queries now work on all columns that have a valid length property, example string and safestr.
Added bracket notation to where statements to let you check against sub propertys, example: ["meta[hairColor]", "LIKE", "blue"].
The bracket notation also works for orderBy and groupBy queries.
Bracket notation also works with .length queries: ["meta[postIDs].length", ">", 2].

[0.8.92] 6-14-2017

Updated to the newest version of Lie-TS, fixing an issue with MS Edge and IE.

[0.8.91] 6-13-2017

Small size improvement.
Added two new config options to the default DB: writeCache and readCache. Lets you adjust those values for LevelDB (in MB).

[0.8.9] 6-9-2017

Merged Pull request #14.

[0.8.82] 6-9-2017

Resolved another issue with complicated where statements.

[0.8.81] 6-6-2017

Fixed an issue with combined where statements.

[0.8.8] 6-3-2017

Made manualExec a public method of query, allowing you to pass the query object itself instead of using .where().join()....
Added * event to trigger on all database events.
Made selected table variable public.
Fixed event handler always showing the event type as "error".

[0.8.72] 6-3-2017

BREAKING CHANGE: describe queries have been reshaped and now also return the primary key and number of rows of the selected table.
Fixed an issue with ORM updates on deletes.

[0.8.71] 6-1-2017

Added NOT HAVE where condition, to find records that don't have a specific value inside an array column.

[0.8.7] 5-31-2017

Fixed an issue with combined where statements.
Fixed an issue causing the event objects not to populate in time.
Slightly optimized full table scans with where statements.

[0.8.63] 5-30-2017

Fixed ORM issue with accidental duplication of references.

[0.8.62] 5-29-2017

Fixed an issue with default row values not being set.
Adjusted the internal _assign method to check if the object is frozen before running JSON operations.

[0.8.61] 5-27-2017

Fixed some ORM issues, the ORM system is a much more stable feature now.
Promise.chain method was not actually chaining methods but running them all at once, causing big problems when sequential processing is needed. Switched to a different chaining method, helping to stabilize the history and ORM system more.

[0.8.6] 5-25-2017

The NanoSQLInstance method random16bits is now public instead of private.
Fixed an issue with the ORM not correctly removing entries.

[0.8.54] 5-23-2017

Fixed an issue where CSV export would skip undefined/false columns entirely, breaking the column alignment.

[0.8.53] 5-22-2017

Fixed an issue where the query would error when using columns containing undfined.

[0.8.52] 5-20-2017

Fixed an issue with compound where statements using primary or secondary keys.
Added optional argument to loadJS and loadCSV to disable transactions for the import.
More history bugfixes.

[0.8.51] 5-19-2017

BREAKING CHANGE: ["orm::..."] model props should be changed to ["ref=>..."].
Fixed a few issues with the history system.
Updated Todo example with different CSS lib.

[0.8.5] 5-17-2017

BREAKING CHANGE: Reworked transaction system to allow parallel transactions. See new API in the documentation.
Restored query memoization.
Issue #11 fully resolved.

[0.8.41] 5-16-2017

Issue #11 partially implimented, history triggered events now return the rows after the history action was performed.
Fixed an issue with the ORM imports with loadJS and loadCSV.

[0.8.4] 5-13-2017

Updated the lib logo and chose a mascot, the hummingbird. 😃
Added a complete ORM system.
.length can now be used with array/orm columns for most queries: select: .query("select",["arrayColumn.length AS Total"]) where: .where(["arrayColumn.length", ">", 2]) groupBy: .groupBy({"arrayColumn.length":"asc"}) orderBy: .orderBy({"arrayColumn.length":"asc"})
Temprarily disabled query memoization to resolve ORM issue. Memoization does not get along with transactions very well.

[0.8.31] 5-11-2017

Fixed an issue preventing falsy like values from being inserted into the database.

[0.8.3] 5-9-2017

Fixed an issue with history not working for the memory only store.

[0.8.21] 4-28-2017

Restored blob and any as valid types.

[0.8.2] 4-28-2017

Integration tests written, all basic SQL functions covered now.
Fixed a variety of small bugs discovered by the tests.
A few small code optimizations.

[0.8.1] 4-24-2017

Added new safestr type, identical to string except it escapes all HTML and unsafe JSON charecters.
Fixed drop/delete bug.
Fixed an issue with callstack exceeded when type casting specific types.

[0.7.92] 4-17-2017

Fixed null values in string columns being type casted to "null".

[0.7.91] 4-13-2017

Fixed secondary indexes not removing old entries.
Moved secondary index updates out of the transaction loop, writes twice as fast now.

[0.7.9] 4-11-2017

Updated promise lib with smaller setImediate polyfill, in browser performance is 10x and lib is only 200 bytes larger.
Moved Trie implimintation to external lib.
Increased leveldown write buffer size.
Added trie to primary key indexes, writes are 10 - 20x faster now.

[0.7.8] 4-11-2017

Fixed an issue with joins.
Tries now rebuild after transactions.
Fixed an issue with tries not pulling results from the data store correctly.

[0.7.75] 4-11-2017

Fixed UglifyJS breaking the history system.
Small size/performance optimizations.

[0.7.7] 4-10-2017

Switched to internal, smaller trie implimentation, reduced gzip size by more than half a kb.
Removed setImediate polyfill, saved another gzip half kb.
Fixed multiple bugs with the history system:

Changes to the same row would sometimes overwrite previous row edits.
History state wouldn't persist correctly into indexedDB/LevelUP
The code that cleans future history points wasn't working correctly.

Improved write speed significantly.
History Points now have an in memory secondary index, increasing the history undo/redo performance significantly.

[0.7.61] 4-9-2017

Trigger change events after transactions now only affect the tables that the transaction touched.
Restored memory db usage.

[0.7.6] 4-8-2017

Secondary indexes and trie indexing now support multiple rows per entry. Existing secondary indexes should be rebuilt with rebuildIndexes:true before using this version.
Added naive event triggering on all tables after a transaction.

[0.7.5] 4-7-2017

Removed transactions from index rebuilding, this was causing the index rebuild to fail on data sets in the thousands.
Removed automatic index sorting to speed up inserts again.
Added a trie implimentation that can be taken advantage of by entering props: ["trie"] into the data model. The trie will always create a secondary index as well meaing props: ["trie", "idx"] is identical to props:["trie"].
Added setImmediate polyfill to increase promise speed, a significant factor contributing to poor performance.

[0.7.4] 4-6-2017

Finished secondary index support. You can now add props: ["idx"] to a data model, when a secondary index or primary key is used in a where statement the data will be retrieved much faster.
Added rebuildIndexes:true config parameter. When passed, the lib will rebuild all secondary indexes before finishing the connect() command. This is mainly to let folks add secondary indexes to existing data models.

[0.7.3] 4-6-2017

Added range() functionality to join() queries.

[0.7.21] 4-6-2017

Added a new .range() query modifier optimized for pagination style queries.
Added a new timeId and timeIdms types that generates a unique, random sortable id with 32 bits of randomness.

[0.7.1] 4-5-2017

Implimented a new TS style array api for data models, the old style is kept around to prevent breaking changes but moving forward types of array should become any[]. You can also typecast the array like bool[] and even nest the arrays like string[][].
Added new any type for data models, can be used with array type casting as well with any and any[].
Switched to setImmediate in NodeJS, write speed is now 8 records/ms, 800 times faster.
Disabled history/event triggering for transactions to increase transaction speed.

[0.7.0] 4-3-2017

Added uuid interface.
Fixed issue with existing indexes not being imported when memory: false and persistent: true.
Removed code that added history meta for every row, even if history is disabled.
Added transaction support to level DB.

[0.6.93] 4-2-2017

Added id config option to default storage driver.

[0.6.92] 4-1-2017

Code optimizations, removed local storage from clearing everything on setup.
Removed global polyfill.

[0.6.91] 3-31-2017

Made cleanArgs method public.
Added primary key range optimizations for Level DB and Indexed DB.
Fixed issue with history when using UUIDs
History is now disabled for tables with no primary key.
Adjusted BETWEEN to be consistent regardless of data store being used.

[0.6.8] 3-30-2017

Added selected table to Action/View filter function.
Added a misc data object to make passing around information easier.

[0.6.6] 3-30-2017

Added filter function for actions & views.
Refactored Actions/views to take less space.
Fixed bug with group by being used with functions.
Adjusted events system a little.

[0.6.5] 3-29-2017

Added HAVE where condition useful for searching arrays in the database.
Changed LIKE behavior where strings will be lowercased to match string values with different casing.
Added error condition if WHERE or HAVING aren't passed any arguments.

[0.6.4] 3-28-2017

Fixed a bug that prevented LevelDB stores from restoring into the memory DB in some situations.
Switched levelDB to JSON encoding.
Fixed clear history bug.

[0.6.3] 3-27-2017

Fixed a bug caused by the primary key optimization that prevented compound where statements from working correctly.

[0.6.2] 3-26-2017

Added function to clear all history
Added primary key opitmization to reads when you do .where(["primaryKey","=",rowID]).
Fixed some issues caused by uglifyJS.
History flush and database flush now implimented completely.

[0.6.1] 3-25-2017

Cleaned up some of the code.
Did some size optimizations.
Fixed issue where resetting the database mode prevented the store from loading.
Adjusted node/browser build. Lib should work when being webpack included AND in node without work from the dev using it.

[0.6.0] 3-25-2017

Fixed an issue with UUIDs
Refactored the abstraction layer and memory store to handle parallel queries.
BREAKING CHANGE: loadCSV and loadJS now require you pass in the table as the first argument, then the data to import
Added query filter method.
Changed the way nodejs packages are being brought in.
Added OPEN open contribution stuff to the ReadMe.
Added license to ReadMe
Fixed issue with NodeJS crypto

[0.5.1] 3-24-2017

Small bugfixes.

[0.5.0] 3-24-2017

BREAKING CHANGE: before_import and after_import have now been switched to .beginTransaction() and .endTransaction() syntax. See the docs.
BREAKING CHANGE: The delete syntax was not deleting entire rows when no arguments were passed as it should, now it is.
config parameters have been added to handle history, immutable functionionality, and persistent storage type.
BREAKING CHANGE: flush_db now deletes in memory store and history as well, not just the persistent storage.
Default store changes:
- Persists history states to IndexedDB or localStorage.
- Falls back to localStorage if IndexedDB isn't availble.
- Persists in NodeJS using LevelDB.
- Rewritten to reduce memory usage dramatically when persist is true and memory is false.
- Using transactions now allow you to chain multiple table updates/changes inside the same history point.
- Added extra functions to the history system to give more control.
- Join commands are twice as fast as before.

[0.4.5] 3-14-2017

BREAKING CHANGE: Did more research on how functions typically work in a SQL envrionment, changed the function implemintation again.
Fixed a bug related to cross joins.

[0.4.4] 3-13-2017

Added BETWEEN condition.
Fixed demo links.

[0.4.3] 3-13-2017

Changed Readme code sections to js to work with NPM better.

[0.4.2] 3-13-2017

Readme changes again.

[0.4.1] 3-13-2017

Readme changes.

[0.4.0] 3-13-2017

Switched (again) to another promise lib, ported from lie.js
Changing lib name to "NanoSQL". Way cooler.
Refactored data store: selects, upserts & joins are now twice as fast.
Now supports tables without primary keys.
Much better conformance to SQL standards.
Added outer joins.
Added more documentation.
Added GroupBy statement.
Added Having statement.
Added AS handling in the select queries.
Removed filter statement, changed filter/function handling.
BREAKING CHANGE: The history pointer is now reversed from it's previous behavior.
BREAKING CHANGE: Filters are now Functions and no longer work like they did before, check the API docs.

[0.3.4] 3-7-2017

ReadMe changes
Fixed event handling bug.

[0.3.3] 3-6-2017

Changed the event handling.
Adjusted examples a bit.

[0.3.2] 3-5-2017

Finished react draw example.
Added error handling for improper connect() usage.

[0.3.1] 3-4-2017

Added a new "blob" row type that bypasses the JSON parsing and freeze functions.

[0.3.0] 3-3-2017

Fixed nodejs behavior if you have indexed db enabled.
Fixed nodejs behavior with UUID crypto.
Some performance and stability improvements.
Switched to a better promise implementation.

[0.2.9] 2-28-2017

Restored JSON.parse(JSON.stringify()) in some places where recursive deep cloning is needed.
Fixed a few small bugs.
Added NodeJS Crypto to the UUID function.

[0.2.8] 2-28-2017

Added "default" behavior to data models.
Added insert filters option to data models.
All upserts are type casted now, regardless of their source.
Upgraded to Typescript 2.2

[0.2.7] 2-24-2017

Fixed iOS bug with indexedDB.

[0.2.6] 2-15-2017

Fixed delete bug in Memory DB.
Fixed event trigger issue with ReactJS.
Cleaned up Readme a bit.

[0.2.5] 2-14-2017

Fixed some readme typos
Added "clear" ability to indexedDB.

[0.2.4] 2-13-2017

Updated & cleaned up documentation a bit.
Changed the few JSON.parse(JSON.stringify()) statements to Object.assign.
Added Indexed DB functionality to memory db.
Upgraded to Webpack 2.
Restored UUID functionality.
Added IndexedDB functionality.

[0.2.3] 2-6-2017

Fixed typo in package.json

[0.2.2] 2-6-2017

Fixed typings reference.

[0.2.1] 2-6-2017

Added Join API.
Adjusted the node build.
Started adding tests.
Completely rewrote the Memory DB with significant performance and memory improvements.
Memory DB now uses hash maps and indexes; no more deep copying.
Memory DB now only returns immutable structures.
Added history API to memory DB for stupid simple version control.
Removed UUID function.
Only marginal increase in lib size.

[0.2.0] 1-28-2017

Some size & speed optimizations.
Fixed orderby bug.

[0.1.9] 1-14-2017

Added strictNullChecks, noImplicitAny, noImplicitReturns, and noImplicitThis to tsconfig.
Made changes to code to remove errors introduced by stricter coding standards.
Removed TSMap from the lib.
Updated strongly typed TSPromise.

[0.1.8] 1-7-2017

Cleaned up README.
Added "db" var to connect promise.
Added extend capability to views and actions.

[0.1.7] 1-3-2017

Cleaned up typings on React Todo example.
Removed some unneeded abstractions in the Memory DB
Updated to new promise implementation.
Updated README to reflect API changes.

[0.1.6] 1-1-2017

Added more typings data.
Made some small changes to the ReadMe and Docs.
Modified memory DB queries to always return an array of objects.
Fixed Action/Views bug if you didn't pass arguments.
Typescript now forces 'this' scoping, added DB argument to queries to resolve it.
Added database variable to events.
Added ReactJS Todo Example.

[0.1.5] 12-31-2016

Added optional "default" value in data models.
Modified memory DB filters to always return an array of objects.

[0.1.4] 12-30-2016

Readme typos Fixed
Added ability for custom functions to be called before the db is connected.
Changed the way the custom arguments are handled to be more dynamic.
Fixed the build so comments make it to the TSD files.

[0.1.3] 12-28-2016

Added a ton of documentation.
Implemented TypeDoc
Cleaned up a bunch of typings.

[0.1.2] 12-28-2016

Fixed a typo in one of the interface declarations.
Cleaned up the readme a bit.
Updated to newest Typescript map lib.

[0.1.1] 12-28-2016

Updated TSPromise to newest version with scoping built in.
Removed the special scoped promise class.
Fixed examples to reflect the new class name to conform to TSLint standards.

[0.1.0] 12-24-2016

First stable release, code is safe to start thinking about using in production environment.

Added TSLint, the project now passes all TSLint checks.
Settled the API down, shouldn't be changing much more moving forward.
Two optimized builds are now generated, one for the browser and another for node.

# Change Log

# 2.0 TODO

# [2.3.6] 08-12-2019

# [2.3.5] 07-23-2019

# [2.3.4] 06-24-2019

# [2.3.3] 5-20-2019

# [2.3.2] 5-16-2019

# [2.3.1] 5-10-2019

# [2.3.0] 5-09-2019

# [2.2.9] 5-03-2019

# [2.2.8] 4-22-2019

# [2.2.7] 4-19-2019

# [2.2.6] 4-19-2019

# [2.2.5] 4-14-2019

# [2.2.4] 4-1-2019

# [2.2.3] 3-25-2019

# [2.2.2] 3-24-2019

# [2.2.1] 3-6-2019

# [2.2.0] 2-17-2019

# [2.1.9] 2-12-2019

# [2.1.8] 2-8-2019

# [2.1.7] 2-7-2019

# [2.1.6] 2-6-2019

# [2.1.5] 2-4-2019

# [2.1.4] 1-31-2018

# [2.1.3] 1-30-2019

# [2.1.2] 1-30-2019

# [2.1.1] 1-24-2018

# [2.1.0] 1-22-2018

# [2.0.9] 1-21-2018

# [2.0.8] 1-21-2018

# [2.0.7] 1-21-2018

# [2.0.6] 1-18-2018

# [2.0.5] 1-17-2018

# [2.0.4] 1-16-2018

# [2.0.3] 1-15-2018

# [2.0.2] 1-15-2018

# [2.0.1] 1-15-2018

# [1.8.1] 1-7-2019

# [1.8.0] 12-16-2018

# [1.7.9] 11-10-2018

# [1.7.8] 10-10-2018

# [1.7.7] 10-10-2018

# [1.7.6] 8-10-2018

# [1.7.5] 8-6-2018

# [1.7.4] 8-1-2018

# [1.7.3] 7-23-2018

# [1.7.2] 7-22-2018

# [1.7.1] 7-19-2018

# [1.7.0] 7-3-2018

# [1.6.9] 6-14-2018

# [1.6.8] 6-13-2018

# [1.6.7] 6-12-2018

# [1.6.6] 6-6-2018

# [1.6.5] 6-5-2018

# [1.6.4] 6-4-2018

# [1.6.3] 5-27-2018

# [1.6.2] 5-24-2018

# [1.6.1] 5-20-2018

# [1.6.0] 5-19-2018

# [1.5.9] 5-18-2018

# [1.5.8] 5-17-2018

# [1.5.7] 5-15-2018

# [1.5.6] 5-14-2018

# [1.5.5] 5-13-2018

# [1.5.4] 5-2-2018

# [1.5.3] 5-1-2018

# [1.5.2] 4-24-2018

# [1.5.1] 4-22-2018

# [1.5.0] 4-18-2018

# [1.4.9] 4-17-2018

# [1.4.8] 4-16-2018

# [1.4.7] 4-15-2018

# [1.4.6] 4-15-2018

# [1.4.5] 4-15-2018

# [1.4.4] 4-14-2018

# [1.4.3] 4-9-2018

# [1.4.2] 4-7-2018

# [1.4.1] 4-5-2018

# [1.4.0] 3-23-2018