[GitHub] spark pull request #16550: [SPARK-19178][SQL] convert string of large number...
Github user asfgit closed the pull request at: https://github.com/apache/spark/pull/16550 --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #16550: [SPARK-19178][SQL] convert string of large number...
Github user jaceklaskowski commented on a diff in the pull request: https://github.com/apache/spark/pull/16550#discussion_r95871812 --- Diff: common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java --- @@ -835,6 +835,187 @@ public UTF8String translate(Map dict) { return fromString(sb.toString()); } + private int getDigit(byte b) { +if (b >= '0' && b <= '9') { + return b - '0'; +} +throw new NumberFormatException(toString()); + } + + /** + * Parses this UTF8String to long. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyLong.parseLong in Hive. + */ + public long toLong() { +if (numBytes == 0) { + throw new NumberFormatException("Empty string!"); --- End diff -- Could you remove `!` from the exception? --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #16550: [SPARK-19178][SQL] convert string of large number...
Github user tejasapatil commented on a diff in the pull request: https://github.com/apache/spark/pull/16550#discussion_r95868053 --- Diff: common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java --- @@ -835,6 +835,187 @@ public UTF8String translate(Map dict) { return fromString(sb.toString()); } + private int getDigit(byte b) { +if (b >= '0' && b <= '9') { + return b - '0'; +} +throw new NumberFormatException(toString()); + } + + /** + * Parses this UTF8String to long. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyLong.parseLong in Hive. + */ + public long toLong() { +if (numBytes == 0) { + throw new NumberFormatException("Empty string!"); +} + +byte b = getByte(0); +final boolean negative = b == '-'; +int offset = 0; +if (negative || b == '+') { + offset++; + if (numBytes == 1) { +throw new NumberFormatException(toString()); + } +} + +final byte separator = '.'; +final int radix = 10; +final long stopValue = Long.MIN_VALUE / radix; +long result = 0; + +while (offset < numBytes) { + b = getByte(offset); + offset++; + if (b == separator) { +// We allow decimals and will return a truncated integral in that case. +// Therefore we won't throw an exception here (checking the fractional +// part happens below.) +break; + } + + int digit = getDigit(b); + // We are going to process the new digit and accumulate the result. However, before doing + // this, if the result is already smaller than the stopValue(Long.MIN_VALUE / 10), then + // result * 10 will definitely be smaller than minValue, and we can stop and throw exception. + if (result < stopValue) { +throw new NumberFormatException(toString()); + } + + result = result * radix - digit; + // Since the previous result is less than or equal to stopValue(Long.MIN_VALUE / 10), we can + // just use `result > 0` to check overflow. If result overflows, we should stop and throw + // exception. + if (result > 0) { +throw new NumberFormatException(toString()); + } +} + +// This is the case when we've encountered a decimal separator. The fractional +// part will not change the number, but we will verify that the fractional part +// is well formed. +while (offset < numBytes) { + if (getDigit(getByte(offset)) == -1) { +throw new NumberFormatException(toString()); + } + offset++; +} + +if (!negative) { + result = -result; + if (result < 0) { +throw new NumberFormatException(toString()); + } +} + +return result; + } + + /** + * Parses this UTF8String to int. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyInt.parseInt in Hive. + */ + public int toInt() { +if (numBytes == 0) { + throw new NumberFormatException("Empty string!"); +} + +byte b = getByte(0); +final boolean negative = b == '-'; +int offset = 0; +if (negative || b == '+') { + offset++; + if (numBytes == 1) { +throw new NumberFormatException(toString()); + } +} + +final byte separator = '.'; +final int radix = 10; +final int stopValue = Integer.MIN_VALUE / radix; +int result = 0; + +while (offset < numBytes) { + b = getByte(offset); + offset++; + if (b == separator) { +// We allow decimals and will return a truncated integral in that case. +// Therefore we won't throw an exception here (checking the fractional +// part happens below.) +break; + } + + int digit = getDigit(b); + // We are going to process the new digit and accumulate the result. However
[GitHub] spark pull request #16550: [SPARK-19178][SQL] convert string of large number...
Github user tejasapatil commented on a diff in the pull request: https://github.com/apache/spark/pull/16550#discussion_r95868295 --- Diff: common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java --- @@ -835,6 +835,187 @@ public UTF8String translate(Map dict) { return fromString(sb.toString()); } + private int getDigit(byte b) { +if (b >= '0' && b <= '9') { + return b - '0'; +} +throw new NumberFormatException(toString()); + } + + /** + * Parses this UTF8String to long. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyLong.parseLong in Hive. + */ + public long toLong() { +if (numBytes == 0) { + throw new NumberFormatException("Empty string!"); +} + +byte b = getByte(0); +final boolean negative = b == '-'; +int offset = 0; +if (negative || b == '+') { + offset++; + if (numBytes == 1) { +throw new NumberFormatException(toString()); + } +} + +final byte separator = '.'; +final int radix = 10; +final long stopValue = Long.MIN_VALUE / radix; +long result = 0; + +while (offset < numBytes) { + b = getByte(offset); + offset++; + if (b == separator) { +// We allow decimals and will return a truncated integral in that case. +// Therefore we won't throw an exception here (checking the fractional +// part happens below.) +break; + } + + int digit = getDigit(b); + // We are going to process the new digit and accumulate the result. However, before doing + // this, if the result is already smaller than the stopValue(Long.MIN_VALUE / 10), then + // result * 10 will definitely be smaller than minValue, and we can stop and throw exception. + if (result < stopValue) { +throw new NumberFormatException(toString()); + } + + result = result * radix - digit; + // Since the previous result is less than or equal to stopValue(Long.MIN_VALUE / 10), we can + // just use `result > 0` to check overflow. If result overflows, we should stop and throw + // exception. + if (result > 0) { +throw new NumberFormatException(toString()); + } +} + +// This is the case when we've encountered a decimal separator. The fractional +// part will not change the number, but we will verify that the fractional part +// is well formed. +while (offset < numBytes) { + if (getDigit(getByte(offset)) == -1) { +throw new NumberFormatException(toString()); + } + offset++; +} + +if (!negative) { + result = -result; + if (result < 0) { +throw new NumberFormatException(toString()); + } +} + +return result; + } + + /** + * Parses this UTF8String to int. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyInt.parseInt in Hive. + */ + public int toInt() { +if (numBytes == 0) { + throw new NumberFormatException("Empty string!"); +} + +byte b = getByte(0); +final boolean negative = b == '-'; +int offset = 0; +if (negative || b == '+') { + offset++; + if (numBytes == 1) { +throw new NumberFormatException(toString()); + } +} + +final byte separator = '.'; +final int radix = 10; +final int stopValue = Integer.MIN_VALUE / radix; +int result = 0; + +while (offset < numBytes) { + b = getByte(offset); + offset++; + if (b == separator) { +// We allow decimals and will return a truncated integral in that case. +// Therefore we won't throw an exception here (checking the fractional +// part happens below.) +break; + } + + int digit = getDigit(b); + // We are going to process the new digit and accumulate the result. However
[GitHub] spark pull request #16550: [SPARK-19178][SQL] convert string of large number...
Github user tejasapatil commented on a diff in the pull request: https://github.com/apache/spark/pull/16550#discussion_r95858028 --- Diff: common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java --- @@ -835,6 +835,187 @@ public UTF8String translate(Map dict) { return fromString(sb.toString()); } + private int getDigit(byte b) { +if (b >= '0' && b <= '9') { + return b - '0'; +} +throw new NumberFormatException(toString()); + } + + /** + * Parses this UTF8String to long. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyLong.parseLong in Hive. --- End diff -- `These codes are mostly copied ...` => `This code is mostly copied ...` --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #16550: [SPARK-19178][SQL] convert string of large number...
Github user tejasapatil commented on a diff in the pull request: https://github.com/apache/spark/pull/16550#discussion_r95869209 --- Diff: common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java --- @@ -835,6 +835,185 @@ public UTF8String translate(Map dict) { return fromString(sb.toString()); } + private int getDigit(byte b) { +if (b >= '0' && b <= '9') { + return b - '0'; +} +throw new NumberFormatException(toString()); + } + + /** + * Parses this UTF8String to long. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyLong.parseLong in Hive. + */ + public long toLong() { +if (numBytes == 0) { + throw new NumberFormatException("Empty string!"); +} + +byte b = getByte(0); +final boolean negative = b == '-'; +int offset = 0; +if (negative || b == '+') { + offset++; + if (numBytes == 1) { +throw new NumberFormatException(toString()); + } +} + +final byte separator = '.'; +final long stopValue = Long.MIN_VALUE / 10; +long result = 0; + +while (offset < numBytes) { + b = getByte(offset); + offset++; + if (b == separator) { +// We allow decimals and will return a truncated integral in that case. +// Therefore we won't throw an exception here (checking the fractional +// part happens below.) +break; + } + + int digit = getDigit(b); + // We are going to process the new digit and accumulate the result. However, before doing + // this, if the result is already smaller than the stopValue(Long.MIN_VALUE / 10), then + // result * 10 will definitely be smaller than minValue, and we can stop and throw exception. + if (result < stopValue) { +throw new NumberFormatException(toString()); + } + + result = result * 10 - digit; + // Since the previous result is less than or equal to stopValue(Long.MIN_VALUE / 10), we can + // just use `result > 0` to check overflow. If result overflows, we should stop and throw + // exception. + if (result > 0) { +throw new NumberFormatException(toString()); + } +} + +// This is the case when we've encountered a decimal separator. The fractional +// part will not change the number, but we will verify that the fractional part +// is well formed. +while (offset < numBytes) { + if (getDigit(getByte(offset)) == -1) { +throw new NumberFormatException(toString()); + } + offset++; +} + +if (!negative) { + result = -result; + if (result < 0) { +throw new NumberFormatException(toString()); + } +} + +return result; + } + + /** + * Parses this UTF8String to int. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyInt.parseInt in Hive. + */ + public int toInt() { --- End diff -- can you add a comment about this ? the only reason it will hurt perf is for int conversion where the result will have to be typecasted from long -> int ? --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #16550: [SPARK-19178][SQL] convert string of large number...
Github user tejasapatil commented on a diff in the pull request: https://github.com/apache/spark/pull/16550#discussion_r95868850 --- Diff: common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java --- @@ -835,6 +835,187 @@ public UTF8String translate(Map dict) { return fromString(sb.toString()); } + private int getDigit(byte b) { +if (b >= '0' && b <= '9') { + return b - '0'; +} +throw new NumberFormatException(toString()); + } + + /** + * Parses this UTF8String to long. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyLong.parseLong in Hive. + */ + public long toLong() { +if (numBytes == 0) { + throw new NumberFormatException("Empty string!"); +} + +byte b = getByte(0); +final boolean negative = b == '-'; +int offset = 0; +if (negative || b == '+') { + offset++; + if (numBytes == 1) { +throw new NumberFormatException(toString()); + } +} + +final byte separator = '.'; +final int radix = 10; +final long stopValue = Long.MIN_VALUE / radix; +long result = 0; + +while (offset < numBytes) { + b = getByte(offset); + offset++; + if (b == separator) { +// We allow decimals and will return a truncated integral in that case. +// Therefore we won't throw an exception here (checking the fractional +// part happens below.) +break; + } + + int digit = getDigit(b); + // We are going to process the new digit and accumulate the result. However, before doing + // this, if the result is already smaller than the stopValue(Long.MIN_VALUE / 10), then + // result * 10 will definitely be smaller than minValue, and we can stop and throw exception. + if (result < stopValue) { +throw new NumberFormatException(toString()); + } + + result = result * radix - digit; + // Since the previous result is less than or equal to stopValue(Long.MIN_VALUE / 10), we can + // just use `result > 0` to check overflow. If result overflows, we should stop and throw + // exception. + if (result > 0) { +throw new NumberFormatException(toString()); + } +} + +// This is the case when we've encountered a decimal separator. The fractional +// part will not change the number, but we will verify that the fractional part +// is well formed. +while (offset < numBytes) { + if (getDigit(getByte(offset)) == -1) { +throw new NumberFormatException(toString()); + } + offset++; +} + +if (!negative) { + result = -result; + if (result < 0) { +throw new NumberFormatException(toString()); + } +} + +return result; + } + + /** + * Parses this UTF8String to int. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyInt.parseInt in Hive. --- End diff -- `These codes are mostly copied ...` => `This code is mostly copied ...` --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #16550: [SPARK-19178][SQL] convert string of large number...
Github user gatorsmile commented on a diff in the pull request: https://github.com/apache/spark/pull/16550#discussion_r95850133 --- Diff: sql/core/src/test/resources/sql-tests/inputs/cast.sql --- @@ -0,0 +1,43 @@ +-- cast string representing a valid fractional number to integral should truncate the number +SELECT CAST('1.23' AS int); +SELECT CAST('1.23' AS long); +SELECT CAST('-4.56' AS int); +SELECT CAST('-4.56' AS long); + +-- cast string which are not numbers to integral should return null +SELECT CAST('abc' AS int); +SELECT CAST('abc' AS long); + +-- cast string representing a very large number to integral should return null +SELECT CAST('1234567890123' AS int); +SELECT CAST('12345678901234567890123' AS long); + +-- cast empty string to integral should return null +SELECT CAST('' AS int); +SELECT CAST('' AS long); + +-- cast null to integral should return null +SELECT CAST(NULL AS int); +SELECT CAST(NULL AS long); + +-- cast invalid decimal string to integral should return null +SELECT CAST('123.a' AS int); +SELECT CAST('123.a' AS long); + +-- '-2147483648' is the smallest int value +SELECT CAST('-2147483648' AS int); +SELECT CAST('-2147483649' AS int); + +-- '2147483647' is the largest int value +SELECT CAST('2147483647' AS int); +SELECT CAST('2147483648' AS int); + +-- '-9223372036854775808' is the smallest long value +SELECT CAST('-9223372036854775808' AS long); +SELECT CAST('-9223372036854775809' AS long); + +-- '9223372036854775807' is the largest long value +SELECT CAST('9223372036854775807' AS long); +SELECT CAST('9223372036854775808' AS long); + +-- TODO: migrate all cast tests here. --- End diff -- + 1 --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #16550: [SPARK-19178][SQL] convert string of large number...
Github user gatorsmile commented on a diff in the pull request: https://github.com/apache/spark/pull/16550#discussion_r95850028 --- Diff: common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java --- @@ -835,6 +835,187 @@ public UTF8String translate(Map dict) { return fromString(sb.toString()); } + private int getDigit(byte b) { +if (b >= '0' && b <= '9') { + return b - '0'; +} +throw new NumberFormatException(toString()); + } + + /** + * Parses this UTF8String to long. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyLong.parseLong in Hive. + */ + public long toLong() { +if (numBytes == 0) { + throw new NumberFormatException("Empty string!"); +} + +byte b = getByte(0); +final boolean negative = b == '-'; +int offset = 0; +if (negative || b == '+') { + offset++; + if (numBytes == 1) { +throw new NumberFormatException(toString()); + } +} + +final byte separator = '.'; +final int radix = 10; +final long stopValue = Long.MIN_VALUE / radix; +long result = 0; + +while (offset < numBytes) { + b = getByte(offset); + offset++; + if (b == separator) { +// We allow decimals and will return a truncated integral in that case. +// Therefore we won't throw an exception here (checking the fractional +// part happens below.) +break; + } + + int digit = getDigit(b); + // We are going to process the new digit and accumulate the result. However, before doing + // this, if the result is already smaller than the stopValue(Long.MIN_VALUE / 10), then + // result * 10 will definitely be smaller than minValue, and we can stop and throw exception. + if (result < stopValue) { +throw new NumberFormatException(toString()); + } + + result = result * radix - digit; + // Since the previous result is less than or equal to stopValue(Long.MIN_VALUE / 10), we can + // just use `result > 0` to check overflow. If result overflows, we should stop and throw + // exception. + if (result > 0) { +throw new NumberFormatException(toString()); + } +} + +// This is the case when we've encountered a decimal separator. The fractional +// part will not change the number, but we will verify that the fractional part +// is well formed. +while (offset < numBytes) { + if (getDigit(getByte(offset)) == -1) { +throw new NumberFormatException(toString()); + } + offset++; +} + +if (!negative) { + result = -result; + if (result < 0) { +throw new NumberFormatException(toString()); + } +} + +return result; + } + + /** + * Parses this UTF8String to int. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyInt.parseInt in Hive. --- End diff -- To the other reviewers, in the latest master of Hive, it is `LazyInteger.parseInt` --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #16550: [SPARK-19178][SQL] convert string of large number...
Github user gatorsmile commented on a diff in the pull request: https://github.com/apache/spark/pull/16550#discussion_r95717360 --- Diff: common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java --- @@ -835,6 +835,185 @@ public UTF8String translate(Map dict) { return fromString(sb.toString()); } + private int getDigit(byte b) { +if (b >= '0' && b <= '9') { + return b - '0'; +} +throw new NumberFormatException(toString()); + } + + /** + * Parses this UTF8String to long. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyLong.parseLong in Hive. + */ + public long toLong() { +if (numBytes == 0) { + throw new NumberFormatException("Empty string!"); +} + +byte b = getByte(0); +final boolean negative = b == '-'; +int offset = 0; +if (negative || b == '+') { + offset++; + if (numBytes == 1) { +throw new NumberFormatException(toString()); + } +} + +final byte separator = '.'; +final long stopValue = Long.MIN_VALUE / 10; +long result = 0; + +while (offset < numBytes) { + b = getByte(offset); + offset++; + if (b == separator) { +// We allow decimals and will return a truncated integral in that case. +// Therefore we won't throw an exception here (checking the fractional +// part happens below.) +break; + } + + int digit = getDigit(b); + // We are going to process the new digit and accumulate the result. However, before doing + // this, if the result is already smaller than the stopValue(Long.MIN_VALUE / 10), then + // result * 10 will definitely be smaller than minValue, and we can stop and throw exception. + if (result < stopValue) { +throw new NumberFormatException(toString()); + } + + result = result * 10 - digit; --- End diff -- Then, we can use `radix` here. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #16550: [SPARK-19178][SQL] convert string of large number...
Github user gatorsmile commented on a diff in the pull request: https://github.com/apache/spark/pull/16550#discussion_r95717275 --- Diff: common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java --- @@ -835,6 +835,185 @@ public UTF8String translate(Map dict) { return fromString(sb.toString()); } + private int getDigit(byte b) { +if (b >= '0' && b <= '9') { + return b - '0'; +} +throw new NumberFormatException(toString()); + } + + /** + * Parses this UTF8String to long. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyLong.parseLong in Hive. + */ + public long toLong() { +if (numBytes == 0) { + throw new NumberFormatException("Empty string!"); +} + +byte b = getByte(0); +final boolean negative = b == '-'; +int offset = 0; +if (negative || b == '+') { + offset++; + if (numBytes == 1) { +throw new NumberFormatException(toString()); + } +} + +final byte separator = '.'; +final long stopValue = Long.MIN_VALUE / 10; --- End diff -- Could we define the constant `10` as a variable `radix`? --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #16550: [SPARK-19178][SQL] convert string of large number...
Github user cloud-fan commented on a diff in the pull request: https://github.com/apache/spark/pull/16550#discussion_r95588520 --- Diff: common/unsafe/src/main/java/org/apache/spark/unsafe/types/UTF8String.java --- @@ -835,6 +835,185 @@ public UTF8String translate(Map dict) { return fromString(sb.toString()); } + private int getDigit(byte b) { +if (b >= '0' && b <= '9') { + return b - '0'; +} +throw new NumberFormatException(toString()); + } + + /** + * Parses this UTF8String to long. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyLong.parseLong in Hive. + */ + public long toLong() { +if (numBytes == 0) { + throw new NumberFormatException("Empty string!"); +} + +byte b = getByte(0); +final boolean negative = b == '-'; +int offset = 0; +if (negative || b == '+') { + offset++; + if (numBytes == 1) { +throw new NumberFormatException(toString()); + } +} + +final byte separator = '.'; +final long stopValue = Long.MIN_VALUE / 10; +long result = 0; + +while (offset < numBytes) { + b = getByte(offset); + offset++; + if (b == separator) { +// We allow decimals and will return a truncated integral in that case. +// Therefore we won't throw an exception here (checking the fractional +// part happens below.) +break; + } + + int digit = getDigit(b); + // We are going to process the new digit and accumulate the result. However, before doing + // this, if the result is already smaller than the stopValue(Long.MIN_VALUE / 10), then + // result * 10 will definitely be smaller than minValue, and we can stop and throw exception. + if (result < stopValue) { +throw new NumberFormatException(toString()); + } + + result = result * 10 - digit; + // Since the previous result is less than or equal to stopValue(Long.MIN_VALUE / 10), we can + // just use `result > 0` to check overflow. If result overflows, we should stop and throw + // exception. + if (result > 0) { +throw new NumberFormatException(toString()); + } +} + +// This is the case when we've encountered a decimal separator. The fractional +// part will not change the number, but we will verify that the fractional part +// is well formed. +while (offset < numBytes) { + if (getDigit(getByte(offset)) == -1) { +throw new NumberFormatException(toString()); + } + offset++; +} + +if (!negative) { + result = -result; + if (result < 0) { +throw new NumberFormatException(toString()); + } +} + +return result; + } + + /** + * Parses this UTF8String to int. + * + * Note that, in this method we accumulate the result in negative format, and convert it to + * positive format at the end, if this string is not started with '-'. This is because min value + * is bigger than max value in digits, e.g. Integer.MAX_VALUE is '2147483647' and + * Integer.MIN_VALUE is '-2147483648'. + * + * These codes are mostly copied from LazyInt.parseInt in Hive. + */ + public int toInt() { --- End diff -- Hive also duplicates the code for parsing to long and int, I'm not sure how to remove the duplication without hurting the performance. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #16550: [SPARK-19178][SQL] convert string of large number...
GitHub user cloud-fan opened a pull request: https://github.com/apache/spark/pull/16550 [SPARK-19178][SQL] convert string of large numbers to int should return null ## What changes were proposed in this pull request? When we convert a string to integral, we will convert that string to `decimal(20, 0)` first, so that we can turn a string with decimal format to truncated integral, e.g. `CAST('1.2' AS int)` will return `1`. However, this brings problems when we convert a string with large numbers to integral, e.g. `CAST('1234567890123' AS int)` will return `1912276171`, while Hive returns null as we expected. This is a long standing bug(seems it was there the first day Spark SQL was created), this PR fixes this bug by adding the native support to convert `UTF8String` to integral. ## How was this patch tested? new regression tests You can merge this pull request into a Git repository by running: $ git pull https://github.com/cloud-fan/spark string-to-int Alternatively you can review and apply these changes as the patch at: https://github.com/apache/spark/pull/16550.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #16550 commit afa7066b394217316789c06db467acf40c74cd28 Author: Wenchen Fan Date: 2017-01-11T08:14:44Z native support for converting UTF8String to integral --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org