anoopj commented on code in PR #17201: URL: https://github.com/apache/iceberg/pull/17201#discussion_r3598678547
########## api/src/main/java/org/apache/iceberg/expressions/InclusiveEvalVisitor.java: ########## @@ -0,0 +1,532 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, + * software distributed under the License is distributed on an + * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + * KIND, either express or implied. See the License for the + * specific language governing permissions and limitations + * under the License. + */ +package org.apache.iceberg.expressions; + +import java.util.Collection; +import java.util.Comparator; +import java.util.Set; +import java.util.stream.Collectors; +import org.apache.iceberg.transforms.Transform; +import org.apache.iceberg.types.Comparators; +import org.apache.iceberg.util.NaNUtil; + +public abstract class InclusiveEvalVisitor extends ExpressionVisitors.BoundVisitor<Boolean> { Review Comment: Should we keep it package private for now? Or is it hard to do it because it's in `expressions` sub package? ########## api/src/main/java/org/apache/iceberg/expressions/InclusiveEvalVisitor.java: ########## @@ -0,0 +1,532 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one + * or more contributor license agreements. See the NOTICE file + * distributed with this work for additional information + * regarding copyright ownership. The ASF licenses this file + * to you under the Apache License, Version 2.0 (the + * "License"); you may not use this file except in compliance + * with the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, + * software distributed under the License is distributed on an + * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY + * KIND, either express or implied. See the License for the + * specific language governing permissions and limitations + * under the License. + */ +package org.apache.iceberg.expressions; + +import java.util.Collection; +import java.util.Comparator; +import java.util.Set; +import java.util.stream.Collectors; +import org.apache.iceberg.transforms.Transform; +import org.apache.iceberg.types.Comparators; +import org.apache.iceberg.util.NaNUtil; + +public abstract class InclusiveEvalVisitor extends ExpressionVisitors.BoundVisitor<Boolean> { + private static final int IN_PREDICATE_LIMIT = 200; + + protected static final boolean ROWS_MIGHT_MATCH = true; + protected static final boolean ROWS_CANNOT_MATCH = false; + + /** Return true if null count is non-zero or is unknown, false if null count is 0. */ + protected abstract boolean mayContainNull(int id); + + /** Return true if null count is known and equal to value count, false otherwise. */ + protected abstract boolean containsNullsOnly(int id); + + /** Return true if NaN count is non-zero or is unknown, false if NaN count is 0. */ + protected abstract boolean mayContainNaN(int id); + + /** Return true if NaN count is known and equal to value count, false otherwise. */ + protected abstract boolean containsNaNsOnly(int id); + + /** Return the lower bound if it is known, or null otherwise. */ + protected <T> T lowerBound(BoundReference<T> ref) { + return null; + } + + /** Return the upper bound if it is known, or null otherwise. */ + protected <T> T upperBound(BoundReference<T> ref) { + return null; + } + + /** Return a variant field's lower bound if it is known, or null otherwise. */ + protected <T> T extractLowerBound(BoundExtract<T> bound) { + return null; + } + + /** Return a variant field's upper bound if it is known, or null otherwise. */ + protected <T> T extractUpperBound(BoundExtract<T> bound) { + return null; + } + + @Override + public Boolean alwaysTrue() { + return ROWS_MIGHT_MATCH; // all rows match + } + + @Override + public Boolean alwaysFalse() { + return ROWS_CANNOT_MATCH; // all rows fail + } + + @Override + public Boolean not(Boolean result) { + return !result; + } + + @Override + public Boolean and(Boolean leftResult, Boolean rightResult) { + return leftResult && rightResult; + } + + @Override + public Boolean or(Boolean leftResult, Boolean rightResult) { + return leftResult || rightResult; + } + + @Override + public <T> Boolean isNull(Bound<T> term) { + // no need to check whether the field is required because binding evaluates that case + // if the column has no null values, the expression cannot match + if (isNonNullPreserving(term)) { + // number of non-nulls is the same as for the ref + int id = term.ref().fieldId(); + if (!mayContainNull(id)) { + return ROWS_CANNOT_MATCH; + } + } + + return ROWS_MIGHT_MATCH; + } + + @Override + public <T> Boolean notNull(Bound<T> term) { + // no need to check whether the field is required because binding evaluates that case + // if the column has no non-null values, the expression cannot match + + // all terms are null preserving. see #isNullPreserving(Bound) + int id = term.ref().fieldId(); + if (containsNullsOnly(id)) { + return ROWS_CANNOT_MATCH; + } + + return ROWS_MIGHT_MATCH; + } + + @Override + public <T> Boolean isNaN(Bound<T> term) { + // when there's no nanCounts information, but we already know the column only contains null, + // it's guaranteed that there's no NaN value + int id = term.ref().fieldId(); + if (containsNullsOnly(id)) { + return ROWS_CANNOT_MATCH; + } + + if (!(term instanceof BoundReference)) { + return ROWS_MIGHT_MATCH; + } + + if (!mayContainNaN(id)) { + return ROWS_CANNOT_MATCH; + } + + return ROWS_MIGHT_MATCH; + } + + @Override + public <T> Boolean notNaN(Bound<T> term) { + if (!(term instanceof BoundReference)) { + // identity transforms are already removed by this time + return ROWS_MIGHT_MATCH; + } + + int id = term.ref().fieldId(); + + if (containsNaNsOnly(id)) { + return ROWS_CANNOT_MATCH; + } + + return ROWS_MIGHT_MATCH; + } + + @Override + public <T> Boolean lt(Bound<T> term, Literal<T> lit) { + // all terms are null preserving. see #isNullPreserving(Bound) + int id = term.ref().fieldId(); + if (containsNullsOnly(id) || containsNaNsOnly(id)) { + return ROWS_CANNOT_MATCH; + } + + T lower = evalLowerBound(term); + if (null == lower || NaNUtil.isNaN(lower)) { + // NaN indicates unreliable bounds. See the InclusiveMetricsEvaluator docs for more. + return ROWS_MIGHT_MATCH; + } + + // this also works for transforms that are order preserving: + // if a transform f is order preserving, a < b means that f(a) <= f(b). + // because lower <= a for all values of a in the file, f(lower) <= f(a). + // when f(lower) >= X then f(a) >= f(lower) >= X, so there is no a such that f(a) < X + // f(lower) >= X means rows cannot match + int cmp = lit.comparator().compare(lower, lit.value()); + if (cmp >= 0) { + return ROWS_CANNOT_MATCH; + } + + return ROWS_MIGHT_MATCH; + } + + @Override + public <T> Boolean ltEq(Bound<T> term, Literal<T> lit) { + // all terms are null preserving. see #isNullPreserving(Bound) + int id = term.ref().fieldId(); + if (containsNullsOnly(id) || containsNaNsOnly(id)) { + return ROWS_CANNOT_MATCH; + } + + T lower = evalLowerBound(term); + if (null == lower || NaNUtil.isNaN(lower)) { + // NaN indicates unreliable bounds. See the InclusiveMetricsEvaluator docs for more. + return ROWS_MIGHT_MATCH; + } + + // this also works for transforms that are order preserving: + // if a transform f is order preserving, a < b means that f(a) <= f(b). + // because lower <= a for all values of a in the file, f(lower) <= f(a). + // when f(lower) > X then f(a) >= f(lower) > X, so there is no a such that f(a) <= X + // f(lower) > X means rows cannot match + int cmp = lit.comparator().compare(lower, lit.value()); + if (cmp > 0) { + return ROWS_CANNOT_MATCH; + } + + return ROWS_MIGHT_MATCH; + } + + @Override + public <T> Boolean gt(Bound<T> term, Literal<T> lit) { + // all terms are null preserving. see #isNullPreserving(Bound) + int id = term.ref().fieldId(); + if (containsNullsOnly(id) || containsNaNsOnly(id)) { + return ROWS_CANNOT_MATCH; + } + + T upper = evalUpperBound(term); + if (null == upper) { + return ROWS_MIGHT_MATCH; + } + + int cmp = lit.comparator().compare(upper, lit.value()); + if (cmp <= 0) { + return ROWS_CANNOT_MATCH; + } + + return ROWS_MIGHT_MATCH; + } + + @Override + public <T> Boolean gtEq(Bound<T> term, Literal<T> lit) { + // all terms are null preserving. see #isNullPreserving(Bound) + int id = term.ref().fieldId(); + if (containsNullsOnly(id) || containsNaNsOnly(id)) { + return ROWS_CANNOT_MATCH; + } + + T upper = evalUpperBound(term); + if (null == upper) { + return ROWS_MIGHT_MATCH; + } + + int cmp = lit.comparator().compare(upper, lit.value()); + if (cmp < 0) { + return ROWS_CANNOT_MATCH; + } + + return ROWS_MIGHT_MATCH; + } + + @Override + public <T> Boolean eq(Bound<T> term, Literal<T> lit) { + // all terms are null preserving. see #isNullPreserving(Bound) + int id = term.ref().fieldId(); + if (containsNullsOnly(id) || containsNaNsOnly(id)) { + return ROWS_CANNOT_MATCH; + } + + T lower = evalLowerBound(term); + if (lower != null && !NaNUtil.isNaN(lower)) { + int cmp = lit.comparator().compare(lower, lit.value()); + if (cmp > 0) { + return ROWS_CANNOT_MATCH; + } + } + + T upper = evalUpperBound(term); + if (null == upper) { + return ROWS_MIGHT_MATCH; + } + + int cmp = lit.comparator().compare(upper, lit.value()); + if (cmp < 0) { + return ROWS_CANNOT_MATCH; + } + + return ROWS_MIGHT_MATCH; + } + + @Override + public <T> Boolean notEq(Bound<T> term, Literal<T> lit) { + // because the bounds are not necessarily a min or max value, this cannot be answered using + // them. notEq(col, X) with (X, Y) doesn't guarantee that X is a value in col. + // However, when min == max and the file has no nulls or NaN values, we can safely prune + // if that value equals the literal. + T value = uniqueValue(term); + + if (value != null && lit.comparator().compare(value, lit.value()) == 0) { + return ROWS_CANNOT_MATCH; + } + + return ROWS_MIGHT_MATCH; + } + + @Override + public <T> Boolean in(Bound<T> term, Set<T> literalSet) { + // all terms are null preserving. see #isNullPreserving(Bound) + int id = term.ref().fieldId(); + if (containsNullsOnly(id) || containsNaNsOnly(id)) { + return ROWS_CANNOT_MATCH; + } + + Collection<T> literals = literalSet; + + if (literals.size() > IN_PREDICATE_LIMIT) { + // skip evaluating the predicate if the number of values is too big + return ROWS_MIGHT_MATCH; + } + + T lower = evalLowerBound(term); + if (null == lower || NaNUtil.isNaN(lower)) { + // NaN indicates unreliable bounds. See the InclusiveMetricsEvaluator docs for more. + return ROWS_MIGHT_MATCH; + } + + T upper = evalUpperBound(term); + if (null == upper) { + return ROWS_MIGHT_MATCH; + } + + Comparator<T> comparator = ((BoundTerm<T>) term).comparator(); + + literals = + literals.stream() + .filter(v -> comparator.compare(lower, v) <= 0) + .collect(Collectors.toList()); + // if all values are less than lower bound, rows cannot match + if (literals.isEmpty()) { + return ROWS_CANNOT_MATCH; + } + + literals = + literals.stream() + .filter(v -> comparator.compare(upper, v) >= 0) + .collect(Collectors.toList()); + // if remaining values are greater than upper bound, rows cannot match + if (literals.isEmpty()) { + return ROWS_CANNOT_MATCH; + } + + return ROWS_MIGHT_MATCH; + } + + @Override + public <T> Boolean notIn(Bound<T> term, Set<T> literalSet) { + // because the bounds are not necessarily a min or max value, this cannot be answered using + // them. notIn(col, {X, ...}) with (X, Y) doesn't guarantee that X is a value in col. + // However, when min == max and the file has no nulls or NaN values, we can safely prune + // if that value is in the exclusion set. + T value = uniqueValue(term); + + if (value != null && literalSet.contains(value)) { + return ROWS_CANNOT_MATCH; + } + + return ROWS_MIGHT_MATCH; + } + + @Override + public <T> Boolean startsWith(Bound<T> term, Literal<T> lit) { + if (term instanceof BoundTransform && !((BoundTransform<?, ?>) term).transform().isIdentity()) { + // truncate must be rewritten in binding. the result is either always or never compatible + return ROWS_MIGHT_MATCH; + } + + int id = term.ref().fieldId(); + if (containsNullsOnly(id)) { + return ROWS_CANNOT_MATCH; + } + + String prefix = (String) lit.value(); + + Comparator<CharSequence> comparator = Comparators.charSequences(); + + CharSequence lower = (CharSequence) evalLowerBound(term); + if (null == lower) { + return ROWS_MIGHT_MATCH; + } + + // truncate lower bound so that its length in bytes is not greater than the length of prefix + int length = Math.min(prefix.length(), lower.length()); + int cmp = comparator.compare(lower.subSequence(0, length), prefix); + if (cmp > 0) { + return ROWS_CANNOT_MATCH; + } + + CharSequence upper = (CharSequence) evalUpperBound(term); + if (null == upper) { + return ROWS_MIGHT_MATCH; + } + + // truncate upper bound so that its length in bytes is not greater than the length of prefix + length = Math.min(prefix.length(), upper.length()); + cmp = comparator.compare(upper.subSequence(0, length), prefix); + if (cmp < 0) { + return ROWS_CANNOT_MATCH; + } + + return ROWS_MIGHT_MATCH; + } + + @Override + public <T> Boolean notStartsWith(Bound<T> term, Literal<T> lit) { + // the only transforms that produce strings are truncate and identity, which work with this + int id = term.ref().fieldId(); + if (mayContainNull(id)) { + return ROWS_MIGHT_MATCH; + } + + String prefix = (String) lit.value(); + + Comparator<CharSequence> comparator = Comparators.charSequences(); + + // notStartsWith will match unless all values must start with the prefix. This happens when + // the lower and upper bounds both start with the prefix. + CharSequence lower = (CharSequence) evalLowerBound(term); + CharSequence upper = (CharSequence) evalUpperBound(term); + if (null == lower || null == upper) { + return ROWS_MIGHT_MATCH; + } + + // if lower is shorter than the prefix then lower doesn't start with the prefix + if (lower.length() < prefix.length()) { + return ROWS_MIGHT_MATCH; + } + + int cmp = comparator.compare(lower.subSequence(0, prefix.length()), prefix); + if (cmp == 0) { + // if upper is shorter than the prefix then upper can't start with the prefix + if (upper.length() < prefix.length()) { + return ROWS_MIGHT_MATCH; + } + + cmp = comparator.compare(upper.subSequence(0, prefix.length()), prefix); + if (cmp == 0) { + // both bounds match the prefix, so all rows must match the prefix and therefore do not + // satisfy the predicate + return ROWS_CANNOT_MATCH; + } + } + + return ROWS_MIGHT_MATCH; + } + + /** + * Returns the column's single value if all rows contain the same value. Defined as a column with + * no nulls, no NaNs, and lower bound equals upper bound. Returns null otherwise. + */ + private <T> T uniqueValue(Bound<T> term) { + int id = term.ref().fieldId(); + if (mayContainNull(id)) { + return null; + } + + T lower = evalLowerBound(term); + T upper = evalUpperBound(term); + + if (lower == null || upper == null || NaNUtil.isNaN(lower) || NaNUtil.isNaN(upper)) { + return null; + } + + if (mayContainNaN(id)) { + return null; Review Comment: There is a subtle behavior change here: the previous version of `uniqueValue` returned `null` only if there are known `NaN` values. ie: ``` if (nanCounts != null && nanCounts.containsKey(id) && nanCounts.get(id) != 0) { return null; } ``` But this version does so if the stats are missing (e.g. a if the column type is `string`. This is likely causing CI failures like below: ``` TestSparkScan > testUnpartitionedYears() > catalogName = testhadoop, implementation = org.apache.iceberg.spark.SparkCatalog, config = {type=hadoop, cache-enabled=false}, format = parquet FAILED java.lang.AssertionError: Expected size: 5 but was: 10 in: [org.apache.iceberg.spark.source.SparkInputPartition@[...] at org.apache.iceberg.spark.source.TestSparkScan.testUnpartitionedYears(TestSparkScan.java:496) ``` Did you mean to make this change? The PR description says refactor. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
