Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Anatomy of a translating compiler
Search
Serge Smertin
June 30, 2022
Programming
0
380
Anatomy of a translating compiler
How to build a programming language in Scala?
Serge Smertin
June 30, 2022
Tweet
Share
More Decks by Serge Smertin
See All by Serge Smertin
Building Databricks integrations with Go
nfx
0
230
Building robust Python applications on top of Databricks: UCX case study
nfx
0
220
Introducing the Power of Databricks SDK
nfx
0
370
Reaching 3M downloads and 90%+ unit test coverage for a Terraform provider: lessons learned
nfx
0
410
Distributed Data Mesh, Delta Lake, and Terraform
nfx
0
440
Data Quality with or without Apache Spark and its ecosystem
nfx
0
530
Data Privacy with Apache Spark: Defensive and Offensive Approaches
nfx
0
500
Other Decks in Programming
See All in Programming
Amazon ECS とマイクロサービスから考えるシステム構成
hiyanger
2
560
GAEログのコスト削減
mot_techtalk
0
120
楽しく向き合う例外対応
okutsu
0
120
2,500万ユーザーを支えるSREチームの6年間のスクラムのカイゼン
honmarkhunt
6
5.3k
チームリードになって変わったこと
isaka1022
0
200
2024年のWebフロントエンドのふりかえりと2025年
sakito
3
250
『GO』アプリ バックエンドサーバのコスト削減
mot_techtalk
0
140
Domain-Driven Transformation
hschwentner
2
1.9k
仕様変更に耐えるための"今の"DRY原則を考える / Rethinking the "Don't repeat yourself" for resilience to specification changes
mkmk884
0
190
Ruby on cygwin 2025-02
fd0
0
140
Writing documentation can be fun with plugin system
okuramasafumi
0
120
Conform を推す - Advocating for Conform
mizoguchicoji
3
690
Featured
See All Featured
Responsive Adventures: Dirty Tricks From The Dark Corners of Front-End
smashingmag
251
21k
How to Think Like a Performance Engineer
csswizardry
22
1.3k
A designer walks into a library…
pauljervisheath
205
24k
What’s in a name? Adding method to the madness
productmarketing
PRO
22
3.3k
Principles of Awesome APIs and How to Build Them.
keavy
126
17k
Automating Front-end Workflow
addyosmani
1368
200k
For a Future-Friendly Web
brad_frost
176
9.5k
Performance Is Good for Brains [We Love Speed 2024]
tammyeverts
7
630
Why Our Code Smells
bkeepers
PRO
336
57k
Statistics for Hackers
jakevdp
797
220k
Building a Scalable Design System with Sketch
lauravandoore
461
33k
The Art of Delivering Value - GDevCon NA Keynote
reverentgeek
10
1.3k
Transcript
Anatomy of a translating compiler … or how to use
debugger all the time. 1 Serge Smertin Senior Resident Solutions Architect Databricks
None
Apache Spark query execution recap
4 code IN(4*, 5*)
5 code IN(4*, 5*)
6 code IN(4*, 5*)
7 SearchCommand( FieldIn("code", Seq( Wildcard("4*"), Wildcard("5*") )))
8 SearchCommand( FieldIn("code", Seq( Wildcard("4*"), Wildcard("5*") ))) Or( Like(UnresolvedAttribute("code"), Literal.create("4%"),
'\\'), Like(UnresolvedAttribute("code"), Literal.create("5%"), '\\'))
9 SearchCommand( FieldIn("code", Seq( Wildcard("4*"), Wildcard("5*") ))) Or( Like(UnresolvedAttribute("code"), Literal.create("4%"),
'\\'), Like(UnresolvedAttribute("code"), Literal.create("5%"), '\\')) private def expression(ctx: LogicalContext , expr: ast.Expr): Expression = expr match { case ast.FieldIn(field, exprs) if exprs.exists(_.isInstanceOf[ast.Wildcard]) => val expand = (expr: ast.Expr) => ast.Binary(ast.Field(field), ast.Equals, expr) expression(ctx, exprs.tail.foldLeft(expand(exprs.head)) { (left, right) => ast.Binary(left, ast.Or, expand(right)) }) case ast.FieldIn(field, exprs) => In(UnresolvedAttribute(field), exprs.map(expression(ctx, _))) case ast.Binary(left, ast.Equals, ast.Wildcard(pattern)) => like(ctx, left, pattern) case ast.Binary(left, symbol, right) => symbol match { case ast.Or => Or(attrOrExpr(ctx, left), attrOrExpr(ctx, right)) // ... } // ... }
10 SearchCommand( FieldIn("code", Seq( Wildcard("4*"), Wildcard("5*") ))) Or( Like(UnresolvedAttribute("code"), Literal.create("4%"),
'\\'), Like(UnresolvedAttribute("code"), Literal.create("5%"), '\\')) display(spark.table('main') .where((F.col('code').like('4%') | F.col('code').like('5%'))))
11 SearchCommand( FieldIn("code", Seq( Wildcard("4*"), Wildcard("5*") ))) Or( Like(UnresolvedAttribute("code"), Literal.create("4%"),
'\\'), Like(UnresolvedAttribute("code"), Literal.create("5%"), '\\')) display(spark.table('main') .where((F.col('code').like('4%') | F.col('code').like('5%')))) case relation: UnresolvedRelation => s"spark.table(${q(relation.name)})" private def unfoldWheres(expr: Expression): String = expr match { case And(left, right) => s"${unfoldWheres(left)}\n${unfoldWheres(right)}" case _ => s".where(${expressionCode(expr)})" } private def expressionCode(expr: Expression): String = expr match { case b: BinaryOperator => val symbol = jvmToPythonOverrides.getOrElse(b.symbol, b.symbol) s"(${expressionCode(b.left)} $symbol ${expressionCode(b.right)})" case attr: UnresolvedAttribute => s"F.col(${q(attr.name)})" case Like(col, Literal(value, _ @ StringType), _) => s"${expressionCode(col)}.like('$value')" case _ => s"F.expr(${q(expr.sql)})" }