|
| 1 | +/* |
| 2 | + * Licensed to the Apache Software Foundation (ASF) under one or more |
| 3 | + * contributor license agreements. See the NOTICE file distributed with |
| 4 | + * this work for additional information regarding copyright ownership. |
| 5 | + * The ASF licenses this file to You under the Apache License, Version 2.0 |
| 6 | + * (the "License"); you may not use this file except in compliance with |
| 7 | + * the License. You may obtain a copy of the License at |
| 8 | + * |
| 9 | + * http://www.apache.org/licenses/LICENSE-2.0 |
| 10 | + * |
| 11 | + * Unless required by applicable law or agreed to in writing, software |
| 12 | + * distributed under the License is distributed on an "AS IS" BASIS, |
| 13 | + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 14 | + * See the License for the specific language governing permissions and |
| 15 | + * limitations under the License. |
| 16 | + */ |
| 17 | + |
| 18 | +package org.apache.spark.unsafe.types |
| 19 | + |
| 20 | +import org.apache.commons.lang3.StringUtils |
| 21 | + |
| 22 | +import org.scalacheck.{Arbitrary, Gen} |
| 23 | +import org.scalatest.prop.GeneratorDrivenPropertyChecks |
| 24 | +// scalastyle:off |
| 25 | +import org.scalatest.{FunSuite, Matchers} |
| 26 | + |
| 27 | +import org.apache.spark.unsafe.types.UTF8String.{fromString => toUTF8} |
| 28 | + |
| 29 | +/** |
| 30 | + * This TestSuite utilize ScalaCheck to generate randomized inputs for UTF8String testing. |
| 31 | + */ |
| 32 | +class UTF8StringPropertyCheckSuite extends FunSuite with GeneratorDrivenPropertyChecks with Matchers { |
| 33 | +// scalastyle:on |
| 34 | + |
| 35 | + test("toString") { |
| 36 | + forAll { (s: String) => |
| 37 | + assert(toUTF8(s).toString() === s) |
| 38 | + } |
| 39 | + } |
| 40 | + |
| 41 | + test("numChars") { |
| 42 | + forAll { (s: String) => |
| 43 | + assert(toUTF8(s).numChars() === s.length) |
| 44 | + } |
| 45 | + } |
| 46 | + |
| 47 | + test("startsWith") { |
| 48 | + forAll { (s: String) => |
| 49 | + val utf8 = toUTF8(s) |
| 50 | + assert(utf8.startsWith(utf8)) |
| 51 | + for (i <- 1 to s.length) { |
| 52 | + assert(utf8.startsWith(toUTF8(s.dropRight(i)))) |
| 53 | + } |
| 54 | + } |
| 55 | + } |
| 56 | + |
| 57 | + test("endsWith") { |
| 58 | + forAll { (s: String) => |
| 59 | + val utf8 = toUTF8(s) |
| 60 | + assert(utf8.endsWith(utf8)) |
| 61 | + for (i <- 1 to s.length) { |
| 62 | + assert(utf8.endsWith(toUTF8(s.drop(i)))) |
| 63 | + } |
| 64 | + } |
| 65 | + } |
| 66 | + |
| 67 | + test("toUpperCase") { |
| 68 | + forAll { (s: String) => |
| 69 | + assert(toUTF8(s).toUpperCase === toUTF8(s.toUpperCase)) |
| 70 | + } |
| 71 | + } |
| 72 | + |
| 73 | + test("toLowerCase") { |
| 74 | + forAll { (s: String) => |
| 75 | + assert(toUTF8(s).toLowerCase === toUTF8(s.toLowerCase)) |
| 76 | + } |
| 77 | + } |
| 78 | + |
| 79 | + test("compare") { |
| 80 | + forAll { (s1: String, s2: String) => |
| 81 | + assert(Math.signum(toUTF8(s1).compareTo(toUTF8(s2))) === Math.signum(s1.compareTo(s2))) |
| 82 | + } |
| 83 | + } |
| 84 | + |
| 85 | + test("substring") { |
| 86 | + forAll { (s: String) => |
| 87 | + for (start <- 0 to s.length; end <- 0 to s.length; if start <= end) { |
| 88 | + assert(toUTF8(s).substring(start, end).toString === s.substring(start, end)) |
| 89 | + } |
| 90 | + } |
| 91 | + } |
| 92 | + |
| 93 | + test("contains") { |
| 94 | + forAll { (s: String) => |
| 95 | + for (start <- 0 to s.length; end <- 0 to s.length; if start <= end) { |
| 96 | + val substring = s.substring(start, end) |
| 97 | + assert(toUTF8(s).contains(toUTF8(substring)) === s.contains(substring)) |
| 98 | + } |
| 99 | + } |
| 100 | + } |
| 101 | + |
| 102 | + val whitespaceChar: Gen[Char] = Gen.choose(0x00, 0x20).map(_.toChar) |
| 103 | + val whitespaceString: Gen[String] = Gen.listOf(whitespaceChar).map(_.mkString) |
| 104 | + val randomString: Gen[String] = Arbitrary.arbString.arbitrary |
| 105 | + |
| 106 | + test("trim, trimLeft, trimRight") { |
| 107 | + // lTrim and rTrim are both modified from java.lang.String.trim |
| 108 | + def lTrim(s: String): String = { |
| 109 | + var st = 0 |
| 110 | + val array: Array[Char] = s.toCharArray |
| 111 | + while ((st < s.length) && (array(st) <= ' ')) { |
| 112 | + st += 1 |
| 113 | + } |
| 114 | + if (st > 0) s.substring(st, s.length) else s |
| 115 | + } |
| 116 | + def rTrim(s: String): String = { |
| 117 | + var len = s.length |
| 118 | + val array: Array[Char] = s.toCharArray |
| 119 | + while ((len > 0) && (array(len - 1) <= ' ')) { |
| 120 | + len -= 1 |
| 121 | + } |
| 122 | + if (len < s.length) s.substring(0, len) else s |
| 123 | + } |
| 124 | + |
| 125 | + forAll( |
| 126 | + whitespaceString, |
| 127 | + randomString, |
| 128 | + whitespaceString |
| 129 | + ) { (start: String, middle: String, end: String) => |
| 130 | + val s = start + middle + end |
| 131 | + assert(toUTF8(s).trim() === toUTF8(s.trim())) |
| 132 | + assert(toUTF8(s).trimLeft() === toUTF8(lTrim(s))) |
| 133 | + assert(toUTF8(s).trimRight() === toUTF8(rTrim(s))) |
| 134 | + } |
| 135 | + } |
| 136 | + |
| 137 | + test("reverse") { |
| 138 | + forAll { (s: String) => |
| 139 | + assert(toUTF8(s).reverse === toUTF8(s.reverse)) |
| 140 | + } |
| 141 | + } |
| 142 | + |
| 143 | + test("indexOf") { |
| 144 | + forAll { (s: String) => |
| 145 | + for (start <- 0 to s.length; end <- 0 to s.length; if start <= end) { |
| 146 | + val substring = s.substring(start, end) |
| 147 | + assert(toUTF8(s).indexOf(toUTF8(substring), 0) === s.indexOf(substring)) |
| 148 | + } |
| 149 | + } |
| 150 | + } |
| 151 | + |
| 152 | + val randomInt = Gen.choose(-100, 100) |
| 153 | + |
| 154 | + test("repeat") { |
| 155 | + def repeat(str: String, times: Int): String = { |
| 156 | + if (times > 0) str * times else "" |
| 157 | + } |
| 158 | + // ScalaCheck always generating too large repeat times which might hang the test forever. |
| 159 | + forAll(randomString, randomInt) { (s: String, times: Int) => |
| 160 | + assert(toUTF8(s).repeat(times) === toUTF8(repeat(s, times))) |
| 161 | + } |
| 162 | + } |
| 163 | + |
| 164 | + test("lpad, rpad") { |
| 165 | + def padding(origin: String, pad: String, length: Int, isLPad: Boolean): String = { |
| 166 | + if (length <= 0) return "" |
| 167 | + if (length <= origin.length) { |
| 168 | + if (length <= 0) "" else origin.substring(0, length) |
| 169 | + } else { |
| 170 | + if (pad.length == 0) return origin |
| 171 | + val toPad = length - origin.length |
| 172 | + val partPad = if (toPad % pad.length == 0) "" else pad.substring(0, toPad % pad.length) |
| 173 | + if (isLPad) { |
| 174 | + pad * (toPad / pad.length) + partPad + origin |
| 175 | + } else { |
| 176 | + origin + pad * (toPad / pad.length) + partPad |
| 177 | + } |
| 178 | + } |
| 179 | + } |
| 180 | + |
| 181 | + forAll ( |
| 182 | + randomString, |
| 183 | + randomString, |
| 184 | + randomInt |
| 185 | + ) { (s: String, pad: String, length: Int) => |
| 186 | + assert(toUTF8(s).lpad(length, toUTF8(pad)) === |
| 187 | + toUTF8(padding(s, pad, length, true))) |
| 188 | + assert(toUTF8(s).rpad(length, toUTF8(pad)) === |
| 189 | + toUTF8(padding(s, pad, length, false))) |
| 190 | + } |
| 191 | + } |
| 192 | + |
| 193 | + val nullalbeSeq = Gen.listOf(Gen.oneOf[String](null: String, randomString)) |
| 194 | + |
| 195 | + test("concat") { |
| 196 | + def concat(orgin: Seq[String]): String = |
| 197 | + if (orgin.exists(_ == null)) null else orgin.mkString |
| 198 | + |
| 199 | + forAll { (inputs: Seq[String]) => |
| 200 | + assert(UTF8String.concat(inputs.map(toUTF8): _*) === toUTF8(inputs.mkString)) |
| 201 | + } |
| 202 | + forAll (nullalbeSeq) { (inputs: Seq[String]) => |
| 203 | + assert(UTF8String.concat(inputs.map(toUTF8): _*) === toUTF8(concat(inputs))) |
| 204 | + } |
| 205 | + } |
| 206 | + |
| 207 | + test("concatWs") { |
| 208 | + def concatWs(sep: String, inputs: Seq[String]): String = { |
| 209 | + if (sep == null) return null |
| 210 | + inputs.filter(_ != null).mkString(sep) |
| 211 | + } |
| 212 | + |
| 213 | + forAll { (sep: String, inputs: Seq[String]) => |
| 214 | + assert(UTF8String.concatWs(toUTF8(sep), inputs.map(toUTF8): _*) === |
| 215 | + toUTF8(inputs.mkString(sep))) |
| 216 | + } |
| 217 | + forAll(randomString, nullalbeSeq) {(sep: String, inputs: Seq[String]) => |
| 218 | + assert(UTF8String.concatWs(toUTF8(sep), inputs.map(toUTF8): _*) === |
| 219 | + toUTF8(concatWs(sep, inputs))) |
| 220 | + } |
| 221 | + } |
| 222 | + |
| 223 | + // TODO: enable this when we find a proper way to generate valid patterns |
| 224 | + ignore("split") { |
| 225 | + forAll { (s: String, pattern: String, limit: Int) => |
| 226 | + assert(toUTF8(s).split(toUTF8(pattern), limit) === |
| 227 | + s.split(pattern, limit).map(toUTF8(_))) |
| 228 | + } |
| 229 | + } |
| 230 | + |
| 231 | + test("levenshteinDistance") { |
| 232 | + forAll { (one: String, another: String) => |
| 233 | + assert(toUTF8(one).levenshteinDistance(toUTF8(another)) === |
| 234 | + StringUtils.getLevenshteinDistance(one, another)) |
| 235 | + } |
| 236 | + } |
| 237 | + |
| 238 | + test("hashCode") { |
| 239 | + forAll { (s: String) => |
| 240 | + assert(toUTF8(s).hashCode() === toUTF8(s).hashCode()) |
| 241 | + } |
| 242 | + } |
| 243 | + |
| 244 | + test("equals") { |
| 245 | + forAll { (one: String, another: String) => |
| 246 | + assert(toUTF8(one).equals(toUTF8(another)) === one.equals(another)) |
| 247 | + } |
| 248 | + } |
| 249 | +} |
0 commit comments