[
https://issues.apache.org/jira/browse/TINKERPOP-3166?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18009347#comment-18009347
]
ASF GitHub Bot commented on TINKERPOP-3166:
-------------------------------------------
Cole-Greer commented on code in PR #3153:
URL: https://github.com/apache/tinkerpop/pull/3153#discussion_r2224003410
##########
gremlin-core/src/main/java/org/apache/tinkerpop/gremlin/util/NumberHelper.java:
##########
@@ -721,6 +723,86 @@ public static Number coerceTo(final Number a, final
Class<? extends Number> claz
return a;
}
+ /**
+ * Casts the given number to the specified numeric type if it can fit into
it.
+ * Otherwise, throw.
+ *
+ * @param a the number to be cast
+ * @param numberToken the number token denoting the desired type to cast
+ * @return the number cast to the specified type
+ * @throws IllegalArgumentException if the specified numeric type is
unsupported
+ * @throws ArithmeticException if the number overflows
+ */
+ public static Number castTo(final Number a, final N numberToken) {
+ Class<?> clazz = numberToken.getType();
+ if (a.getClass().equals(clazz)) {
+ return a;
+ } else if (clazz.equals(Integer.class)) {
+ Long val = getLong(a, numberToken);
+ if (val >= Integer.MIN_VALUE && val <= Integer.MAX_VALUE) {
+ return a.intValue();
+ }
+ } else if (clazz.equals(Long.class)) {
+ return getLong(a, numberToken);
+ } else if (clazz.equals(Float.class)) {
+ // BigDecimal to double will overflow into Infinity, we want to
throw instead of passing through
+ if (!a.getClass().equals(BigDecimal.class) &&
Review Comment:
BigInteger will also be capable of overflowing into Infinity and will need
special handling here as well
##########
gremlin-core/src/main/java/org/apache/tinkerpop/gremlin/util/NumberHelper.java:
##########
@@ -721,6 +723,86 @@ public static Number coerceTo(final Number a, final
Class<? extends Number> claz
return a;
}
+ /**
+ * Casts the given number to the specified numeric type if it can fit into
it.
+ * Otherwise, throw.
+ *
+ * @param a the number to be cast
+ * @param numberToken the number token denoting the desired type to cast
+ * @return the number cast to the specified type
+ * @throws IllegalArgumentException if the specified numeric type is
unsupported
+ * @throws ArithmeticException if the number overflows
+ */
+ public static Number castTo(final Number a, final N numberToken) {
+ Class<?> clazz = numberToken.getType();
+ if (a.getClass().equals(clazz)) {
+ return a;
+ } else if (clazz.equals(Integer.class)) {
+ Long val = getLong(a, numberToken);
+ if (val >= Integer.MIN_VALUE && val <= Integer.MAX_VALUE) {
+ return a.intValue();
+ }
+ } else if (clazz.equals(Long.class)) {
+ return getLong(a, numberToken);
+ } else if (clazz.equals(Float.class)) {
+ // BigDecimal to double will overflow into Infinity, we want to
throw instead of passing through
+ if (!a.getClass().equals(BigDecimal.class) &&
+ (Double.isInfinite(a.doubleValue()) ||
Double.isNaN(a.doubleValue()))) {
+ return a.floatValue();
+ }
+ if (a.doubleValue() >= -Float.MAX_VALUE && a.doubleValue() <=
Float.MAX_VALUE) {
+ return a.floatValue();
+ }
+ } else if (clazz.equals(Double.class)) {
+ // BigDecimal to double will overflow into Infinity, we want to
throw instead of passing through
+ if (!a.getClass().equals(BigDecimal.class) &&
+ (Double.isInfinite(a.doubleValue()) ||
Double.isNaN(a.doubleValue()))) {
+ return a.doubleValue();
+ }
+ if (!Double.isInfinite(a.doubleValue())) {
+ // float losses precision, use string intermediate
+ return a.getClass().equals(Float.class) ?
Double.parseDouble(a.toString()) : a.doubleValue();
Review Comment:
Is there loss of precision when directly using
[Float.doubleValue()](https://docs.oracle.com/javase/8/docs/api/java/lang/Float.html#doubleValue--)?
##########
gremlin-core/src/main/java/org/apache/tinkerpop/gremlin/util/NumberHelper.java:
##########
@@ -721,6 +723,86 @@ public static Number coerceTo(final Number a, final
Class<? extends Number> claz
return a;
}
+ /**
+ * Casts the given number to the specified numeric type if it can fit into
it.
+ * Otherwise, throw.
+ *
+ * @param a the number to be cast
+ * @param numberToken the number token denoting the desired type to cast
+ * @return the number cast to the specified type
+ * @throws IllegalArgumentException if the specified numeric type is
unsupported
+ * @throws ArithmeticException if the number overflows
+ */
+ public static Number castTo(final Number a, final N numberToken) {
+ Class<?> clazz = numberToken.getType();
+ if (a.getClass().equals(clazz)) {
+ return a;
+ } else if (clazz.equals(Integer.class)) {
+ Long val = getLong(a, numberToken);
+ if (val >= Integer.MIN_VALUE && val <= Integer.MAX_VALUE) {
+ return a.intValue();
+ }
+ } else if (clazz.equals(Long.class)) {
+ return getLong(a, numberToken);
+ } else if (clazz.equals(Float.class)) {
+ // BigDecimal to double will overflow into Infinity, we want to
throw instead of passing through
+ if (!a.getClass().equals(BigDecimal.class) &&
+ (Double.isInfinite(a.doubleValue()) ||
Double.isNaN(a.doubleValue()))) {
+ return a.floatValue();
+ }
+ if (a.doubleValue() >= -Float.MAX_VALUE && a.doubleValue() <=
Float.MAX_VALUE) {
+ return a.floatValue();
+ }
+ } else if (clazz.equals(Double.class)) {
+ // BigDecimal to double will overflow into Infinity, we want to
throw instead of passing through
Review Comment:
BigInteger may also overflow here
##########
gremlin-core/src/main/java/org/apache/tinkerpop/gremlin/util/NumberHelper.java:
##########
@@ -721,6 +723,86 @@ public static Number coerceTo(final Number a, final
Class<? extends Number> claz
return a;
}
+ /**
+ * Casts the given number to the specified numeric type if it can fit into
it.
+ * Otherwise, throw.
+ *
+ * @param a the number to be cast
+ * @param numberToken the number token denoting the desired type to cast
+ * @return the number cast to the specified type
+ * @throws IllegalArgumentException if the specified numeric type is
unsupported
+ * @throws ArithmeticException if the number overflows
+ */
+ public static Number castTo(final Number a, final N numberToken) {
+ Class<?> clazz = numberToken.getType();
+ if (a.getClass().equals(clazz)) {
+ return a;
+ } else if (clazz.equals(Integer.class)) {
+ Long val = getLong(a, numberToken);
+ if (val >= Integer.MIN_VALUE && val <= Integer.MAX_VALUE) {
+ return a.intValue();
+ }
+ } else if (clazz.equals(Long.class)) {
+ return getLong(a, numberToken);
+ } else if (clazz.equals(Float.class)) {
+ // BigDecimal to double will overflow into Infinity, we want to
throw instead of passing through
+ if (!a.getClass().equals(BigDecimal.class) &&
+ (Double.isInfinite(a.doubleValue()) ||
Double.isNaN(a.doubleValue()))) {
Review Comment:
I believe this will result in something like this:
```
g.inject(5E50).asNumber(N.nfloat)
===> Infinity (float)
```
as 5E50 can be represented as a non-infinite double, but not a float. Would
we also want to throw in these cases?
> Add number conversion step asNumber()
> -------------------------------------
>
> Key: TINKERPOP-3166
> URL: https://issues.apache.org/jira/browse/TINKERPOP-3166
> Project: TinkerPop
> Issue Type: Improvement
> Components: language
> Affects Versions: 3.8.0
> Reporter: Yang Xia
> Priority: Major
>
> Given the addition of the {{asString()}} and {{asDate()}} steps in the 3.7
> line, it should also be helpful to add an {{asNumber()}} step that does
> numerical casting/conversions.
> The current idea is for the {{asNumber()}} step to convert the incoming
> traverser to the nearest parsable type (e.g. int or double) if no argument is
> provided, or to the desired numerical type, based on a number token
> ({{{}N{}}}) provided. Like the {{asDate()}} step, it will not be scoped (for
> now, scopes can be added in the future).
> Some conjured examples:
> {code:java}
> gremlin> g.inject(5).asNumber()
> ==> 5 // parses to int
> gremlin> g.inject(5.123f).asNumber()
> ==> 5.123
> gremlin> g.inject(5.43).asNumber(N.int)
> ==> 5 {code}
> More details can be found in the [proposal
> doc|https://github.com/apache/tinkerpop/blob/master/docs/src/dev/future/proposal-asnumber-step-6.asciidoc].
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
