1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428
#![warn(missing_docs)] #![crate_name="joinkit"] //! Joinkit provides iterator adaptors for efficient SQL-like joins. //! //! # Strategies //! //! There are two join strategies, which fit different scenarios: //! - **Hash Join** - a shorter data stream is loaded entirely into memory (`HashMap`), while the //! longer can be arbitrarily large and is matched against `HashMap` sequentially. The greatest //! advantage is that data do not need to be sorted and it has amortized O(n) complexity, therefore //! it is very efficient. This is the right choice if data is not sorted and the smaller stream //! fits into memory. //! - **Merge Join** - the data streams *must* be sorted, but can be *both* arbitrarily large. This //! is the right choice if the data is already sorted, as in this case it is slightly more //! efficient than Hash Join. //! //! To use the iterator adaptors in this crate, import `Joinkit trait`: //! //! ``` //! use joinkit::Joinkit; //! ``` //! //! The crate contains also 2 binaries `hjoin` and `mjoin`, which can be used to perform `Hash //! Join` and `Merge Join` on command line. #[macro_use] extern crate clap; extern crate itertools; use std::iter::{IntoIterator}; use std::cmp::Ordering; use std::hash::Hash; pub use merge_join::{MergeJoinInner, MergeJoinLeftExcl, MergeJoinLeftOuter, MergeJoinFullOuter}; pub use hash_join::{HashJoinInner, HashJoinLeftExcl, HashJoinLeftOuter, HashJoinRightExcl, HashJoinRightOuter, HashJoinFullOuter}; pub mod util; mod merge_join; mod hash_join; /// A value yielded by `merge_join` and `hash_join` outer iterators. /// Contains one or two values, depending on which input iterator is exhausted. /// #[derive(Clone, PartialEq, Eq, Debug)] pub enum EitherOrBoth<L, R> { /// Neither input iterator is exhausted yet, yielding two values. Both(L, R), /// The parameter iterator is exhausted, only yielding a value from the `self` iterator. Left(L), /// The `self` iterator is exhausted, only yielding a value from the parameter iterator. Right(R), } /// Trait `Joinkit` provides the extra iterator adaptors for efficient SQL-like joins. pub trait Joinkit : Iterator { /// Return an iterator adaptor that [inner /// joins](https://en.wikipedia.org/wiki/Join_%28SQL%29#Inner_join) the two input iterators in /// ascending order. The resulting iterator is the intersection of the two input iterators. /// /// The both input iterators must be sorted and unique on the join key (e.g. by /// [grouping](http://bluss.github.io/rust-itertools/doc/itertools/trait.Itertools.html#method.group_by) /// them, if necessary) to produce the correct results. /// /// Iterator element type is `(L::Item, R::Item)`. /// /// ``` /// use joinkit::Joinkit; /// /// // tuples of (key, [value,...]), where the key is extracted from the value /// // notice the values are grouped by the key /// let l = vec![("0", vec!["0;A"]), ("1", vec!["1;B"])].into_iter(); /// let r = vec![("1", vec!["1;X", "1;Y"]), ("2", vec!["2;Z"])].into_iter(); /// let mut it = l.merge_join_inner_by(r, |x, y| Ord::cmp(&x.0, &y.0)); /// /// assert_eq!(it.next(), Some((("1", vec!["1;B"]), ("1", vec!["1;X", "1;Y"])))); /// assert_eq!(it.next(), None); /// ``` fn merge_join_inner_by<R, F>(self, other: R, cmp: F) -> MergeJoinInner<Self, R::IntoIter, F> where Self: Sized, R: IntoIterator, F: FnMut(&Self::Item, &R::Item) -> Ordering { MergeJoinInner::new(self, other.into_iter(), cmp) } /// Return an iterator adaptor that *left exclusive joins* the two input iterators in /// ascending order. The resulting iterator contains only those records from the left input /// iterator, which do not match the right input iterator. There is no direct equivalent in /// SQL. /// /// The both input iterators must be sorted and unique on the join key (e.g. by /// [grouping](http://bluss.github.io/rust-itertools/doc/itertools/trait.Itertools.html#method.group_by) /// them, if necessary) to produce the correct results. /// /// Iterator element type is `L::Item`. /// /// ``` /// use joinkit::Joinkit; /// /// // tuples of (key, [value,...]), where the key is extracted from the value /// // notice the values are grouped by the key /// let l = vec![("0", vec!["0;A"]), ("1", vec!["1;B"])].into_iter(); /// let r = vec![("1", vec!["1;X", "1;Y"]), ("2", vec!["2;Z"])].into_iter(); /// let mut it = l.merge_join_left_excl_by(r, |x, y| Ord::cmp(&x.0, &y.0)); /// /// assert_eq!(it.next(), Some(("0", vec!["0;A"]))); /// assert_eq!(it.next(), None); /// ``` fn merge_join_left_excl_by<R, F>(self, other: R, cmp: F) -> MergeJoinLeftExcl<Self, R::IntoIter, F> where Self: Sized, R: IntoIterator, F: FnMut(&Self::Item, &R::Item) -> Ordering { MergeJoinLeftExcl::new(self, other.into_iter(), cmp) } /// Return an iterator adaptor that [left outer /// joins](https://en.wikipedia.org/wiki/Join_%28SQL%29#Left_outer_join) the two input iterators /// in ascending order. The resulting iterator contains all the records from the left input /// iterator, even if they do not match the right input iterator. /// /// The both input iterators must be sorted and unique on the join key (e.g. by /// [grouping](http://bluss.github.io/rust-itertools/doc/itertools/trait.Itertools.html#method.group_by) /// them, if necessary) to produce the correct results. /// /// Iterator element type is [`EitherOrBoth<L::Item, R::Item>`](enum.EitherOrBoth.html). /// /// ``` /// use joinkit::Joinkit; /// use joinkit::EitherOrBoth::{Left, Both, Right}; /// /// // tuples of (key, [value,...]), where the key is extracted from the value /// // notice the values are grouped by the key /// let l = vec![("0", vec!["0;A"]), ("1", vec!["1;B"])].into_iter(); /// let r = vec![("1", vec!["1;X", "1;Y"]), ("2", vec!["2;Z"])].into_iter(); /// let mut it = l.merge_join_left_outer_by(r, |x, y| Ord::cmp(&x.0, &y.0)); /// /// assert_eq!(it.next(), Some(Left(("0", vec!["0;A"])))); /// assert_eq!(it.next(), Some(Both(("1", vec!["1;B"]), ("1", vec!["1;X", "1;Y"])))); /// assert_eq!(it.next(), None); /// ``` fn merge_join_left_outer_by<R, F>(self, other: R, cmp: F) -> MergeJoinLeftOuter<Self, R::IntoIter, F> where Self: Sized, R: IntoIterator, F: FnMut(&Self::Item, &R::Item) -> Ordering { MergeJoinLeftOuter::new(self, other.into_iter(), cmp) } /// Return an iterator adaptor that [full outer /// joins](https://en.wikipedia.org/wiki/Join_%28SQL%29#Full_outer_join) the two input iterators /// in ascending order. The resulting iterator contains all the records from the both input /// iterators. /// /// The both input iterators must be sorted and unique on the join key (e.g. by /// [grouping](http://bluss.github.io/rust-itertools/doc/itertools/trait.Itertools.html#method.group_by) /// them, if necessary) to produce the correct results. /// /// Iterator element type is [`EitherOrBoth<L::Item, R::Item>`](enum.EitherOrBoth.html). /// /// ``` /// use joinkit::Joinkit; /// use joinkit::EitherOrBoth::{Left, Both, Right}; /// /// /// // tuples of (key, [value,...]), where the key is extracted from the value /// // notice the values are grouped by the key /// let l = vec![("0",vec!["0;A"]), ("1", vec!["1;B"])].into_iter(); /// let r = vec![("1",vec!["1;X", "1;Y"]), ("2", vec!["2;Z"])].into_iter(); /// let mut it = l.merge_join_full_outer_by(r, |x, y| Ord::cmp(&x.0, &y.0)); /// /// assert_eq!(it.next(), Some(Left(("0", vec!["0;A"])))); /// assert_eq!(it.next(), Some(Both(("1", vec!["1;B"]), ("1", vec!["1;X", "1;Y"])))); /// assert_eq!(it.next(), Some(Right(("2", vec!["2;Z"])))); /// assert_eq!(it.next(), None); /// ``` fn merge_join_full_outer_by<R, F>(self, other: R, cmp: F) -> MergeJoinFullOuter<Self, R::IntoIter, F> where Self: Sized, R: IntoIterator, F: FnMut(&Self::Item, &R::Item) -> Ordering { MergeJoinFullOuter::new(self, other.into_iter(), cmp) } /// Return an iterator adaptor that [inner /// joins](https://en.wikipedia.org/wiki/Join_%28SQL%29#Inner_join) the two input iterators in /// ascending order. The resulting iterator is the intersection of the two input iterators. /// /// The input iterators do *not* need to be sorted. The right input iterator is loaded into /// `HashMap` and grouped by the key automatically. Neither the left input iterator need to be /// unique on the key. /// /// The left input iterator element type must be `(K, LV)`, where `K: Hash + Eq`. /// The right input iterator element type must be `(K, RV)`, where `K: Hash + Eq` and `RV: /// Clone`. /// /// When the join adaptor is created, the right iterator is **consumed** into `HashMap`. /// /// Iterator element type is `(LV, vec![RV,...])`. /// The `RV` is cloned from `HashMap` for each joined value. A single `RV` can be expected to /// be joined (and cloned) multiple times to `LV`. To increase performance, consider wrapping /// `RV` into `std::rc::Rc` pointer to avoid unnecessary allocations. /// /// ``` /// use joinkit::Joinkit; /// /// // tuples of (key, value), where the key is extracted from the value /// let l = vec![("0", "0;A"), ("1", "1;B")].into_iter(); /// let r = vec![("1", "1;X"), ("2", "2;Z"), ("1", "1;Y")].into_iter(); /// let mut it = l.hash_join_inner(r); /// /// // notice the grouped right values /// assert_eq!(it.next(), Some(("1;B", vec!["1;X", "1;Y"]))); /// assert_eq!(it.next(), None); /// ``` fn hash_join_inner<K, RI, RV>(self, other: RI) -> HashJoinInner<Self, K, RV> where Self: Sized, K: Hash + Eq, RV: Clone, RI: IntoIterator<Item=(K, RV)> { HashJoinInner::new(self, other) } /// Return an iterator adaptor that *left exclusive joins* the two input iterators. The /// resulting iterator contains only those records from the left input iterator, which do not /// match the right input iterator. There is no direct equivalent in SQL. /// /// The input iterators do *not* need to be sorted. The right input iterator is loaded into /// `HashMap` and grouped by the key automatically. Neither the left input iterator need to be /// unique on the key. /// /// The left input iterator element type must be `(K, LV)`, where `K: Hash + Eq`. /// The right input iterator element type must be `(K, RV)`, where `K: Hash + Eq`. /// /// When the join adaptor is created, the right iterator is **consumed** into `HashMap`. /// /// Iterator element type is `LV`. /// /// ``` /// use joinkit::Joinkit; /// /// // tuples of (key, value), where the key is extracted from the value /// let l = vec![("0", "0;A"), ("1", "1;B")].into_iter(); /// let r = vec![("1", "1;X"), ("2", "2;Z"), ("1", "1;Y")].into_iter(); /// let mut it = l.hash_join_left_excl(r); /// /// assert_eq!(it.next(), Some("0;A")); /// assert_eq!(it.next(), None); /// ``` fn hash_join_left_excl<K, RI, RV>(self, other: RI) -> HashJoinLeftExcl<Self, K> where Self: Sized, K: Hash + Eq, RI: IntoIterator<Item=(K, RV)> { HashJoinLeftExcl::new(self, other) } /// Return an iterator adaptor that [left outer /// joins](https://en.wikipedia.org/wiki/Join_%28SQL%29#Left_outer_join) the two input /// iterators. The resulting iterator contains all the records from the left input iterator, /// even if they do not match the right input iterator. /// /// The input iterators do *not* need to be sorted. The right input iterator is loaded into /// `HashMap` and grouped by the key automatically. Neither the left input iterator need to be /// unique on the key. /// /// The left input iterator element type must be `(K, LV)`, where `K: Hash + Eq`. /// The right input iterator element type must be `(K, RV)`, where `K: Hash + Eq` and `RV: /// Clone`. /// /// When the join adaptor is created, the right iterator is **consumed** into `HashMap`. /// /// Iterator element type is [`EitherOrBoth<LV, RV>`](enum.EitherOrBoth.html). /// The `RV` is cloned from `HashMap` for each joined value. It is expected a single `RV` will /// be joined (and cloned) multiple times to `LV`. To increase performance, consider wrapping /// `RV` into `std::rc::Rc` pointer to avoid unnecessary allocations. /// /// ``` /// use joinkit::Joinkit; /// use joinkit::EitherOrBoth::{Left, Both, Right}; /// /// // tuples of (key, value), where the key is extracted from the value /// let l = vec![("0", "0;A"), ("1", "1;B")].into_iter(); /// let r = vec![("1", "1;X"), ("2", "2;Z"), ("1", "1;Y")].into_iter(); /// let mut it = l.hash_join_left_outer(r); /// /// // notice the grouped right values /// assert_eq!(it.next(), Some(Left("0;A"))); /// assert_eq!(it.next(), Some(Both("1;B", vec!["1;X", "1;Y"]))); /// assert_eq!(it.next(), None); /// ``` fn hash_join_left_outer<K, RI, RV>(self, other: RI) -> HashJoinLeftOuter<Self, K, RV> where Self: Sized, K: Hash + Eq, RV: Clone, RI: IntoIterator<Item=(K, RV)> { HashJoinLeftOuter::new(self, other) } /// Return an iterator adaptor that *right exclusive joins* the two input iterators. The resulting /// iterator contains only those records from the right input iterator, which do not match the /// left input iterator. There is no direct equivalent in SQL. /// /// The input iterators do *not* need to be sorted. The right input iterator is loaded into /// `HashMap` and grouped by the key automatically. Neither the left input iterator need to be /// unique on the key. /// /// The left input iterator element type must be `(K, LV)`, where `K: Hash + Eq`. /// The right input iterator element type must be `(K, RV)`, where `K: Hash + Eq`. /// /// When the join adaptor is created, the right iterator is **consumed** into `HashMap`. /// /// Iterator element type is `vec![RV,...]`. /// /// ``` /// use joinkit::Joinkit; /// /// // tuples of (key, value), where the key is extracted from the value /// let l = vec![("0", "0;A"), ("1", "1;B")].into_iter(); /// let r = vec![("1", "1;X"), ("2", "2;Z"), ("1", "1;Y")].into_iter(); /// let mut it = l.hash_join_right_excl(r); /// /// assert_eq!(it.next(), Some(vec!["2;Z"])); /// assert_eq!(it.next(), None); /// ``` fn hash_join_right_excl<K, RI, RV>(self, other: RI) -> HashJoinRightExcl<Self, K, RV> where Self: Sized, K: Hash + Eq, RI: IntoIterator<Item=(K, RV)> { HashJoinRightExcl::new(self, other) } /// Return an iterator adaptor that [right outer /// joins](https://en.wikipedia.org/wiki/Join_%28SQL%29#Right_outer_join) the two input /// iterators. The resulting iterator contains all the records from the right input iterator, /// even if they do not match the left input iterator. /// /// The input iterators do *not* need to be sorted. The right input iterator is loaded into /// `HashMap` and grouped by the key automatically. Neither the left input iterator need to be /// unique on the key. /// /// The left input iterator element type must be `(K, LV)`, where `K: Hash + Eq`. /// The right input iterator element type must be `(K, RV)`, where `K: Hash + Eq` and `RV: /// Clone`. /// /// When the join adaptor is created, the right iterator is **consumed** into `HashMap`. /// /// Iterator element type is [`EitherOrBoth<LV, RV>`](enum.EitherOrBoth.html). /// The `RV` is cloned from `HashMap` for each joined value. It is expected a single `RV` will /// be joined (and cloned) multiple times to `LV`. To increase performance, consider wrapping /// `RV` into `std::rc::Rc` pointer to avoid unnecessary allocations. /// /// ``` /// use joinkit::Joinkit; /// use joinkit::EitherOrBoth::{Left, Both, Right}; /// /// // tuples of (key, value), where the key is extracted from the value /// let l = vec![("0", "0;A"), ("1", "1;B")].into_iter(); /// let r = vec![("1", "1;X"), ("2", "2;Z"), ("1", "1;Y")].into_iter(); /// let mut it = l.hash_join_right_outer(r); /// /// // notice the grouped right values /// assert_eq!(it.next(), Some(Both("1;B", vec!["1;X", "1;Y"]))); /// assert_eq!(it.next(), Some(Right(vec!["2;Z"]))); /// assert_eq!(it.next(), None); /// ``` fn hash_join_right_outer<K, RI, RV>(self, other: RI) -> HashJoinRightOuter<Self, K, RV> where Self: Sized, K: Hash + Eq, RV: Clone, RI: IntoIterator<Item=(K, RV)> { HashJoinRightOuter::new(self, other) } /// Return an iterator adaptor that [full outer /// joins](https://en.wikipedia.org/wiki/Join_%28SQL%29#Full_outer_join) the two input /// iterators. The resulting iterator contains all the records from the both input iterators. /// /// The input iterators do *not* need to be sorted. The right input iterator is loaded into /// `HashMap` and grouped by the key automatically. Neither the left input iterator need to be /// unique on the key. /// /// The left input iterator element type must be `(K, LV)`, where `K: Hash + Eq`. /// The right input iterator element type must be `(K, RV)`, where `K: Hash + Eq` and `RV: /// Clone`. /// /// When the join adaptor is created, the right iterator is **consumed** into `HashMap`. /// /// Iterator element type is [`EitherOrBoth<LV, RV>`](enum.EitherOrBoth.html). /// The `RV` is cloned from `HashMap` for each joined value. It is expected a single `RV` will /// be joined (and cloned) multiple times to `LV`. To increase performance, consider wrapping /// `RV` into `std::rc::Rc` pointer to avoid unnecessary allocations. /// /// ``` /// use joinkit::Joinkit; /// use joinkit::EitherOrBoth::{Left, Both, Right}; /// /// // tuples of (key, value), where the key is extracted from the value /// let l = vec![("0", "0;A"), ("1", "1;B")].into_iter(); /// let r = vec![("1", "1;X"), ("2", "2;Z"), ("1", "1;Y")].into_iter(); /// let mut it = l.hash_join_full_outer(r); /// /// // notice the grouped right values /// assert_eq!(it.next(), Some(Left("0;A"))); /// assert_eq!(it.next(), Some(Both("1;B", vec!["1;X", "1;Y"]))); /// assert_eq!(it.next(), Some(Right(vec!["2;Z"]))); /// assert_eq!(it.next(), None); /// ``` fn hash_join_full_outer<K, RI, RV>(self, other: RI) -> HashJoinFullOuter<Self, K, RV> where Self: Sized, K: Hash + Eq, RV: Clone, RI: IntoIterator<Item=(K, RV)> { HashJoinFullOuter::new(self, other) } } impl<T: ?Sized> Joinkit for T where T: Iterator { }