suchagit commented on a change in pull request #9356: URL: https://github.com/apache/arrow/pull/9356#discussion_r585730134
########## File path: csharp/src/Apache.Arrow/DecimalUtility.cs ########## @@ -0,0 +1,163 @@ +// 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. + +using System; +using System.Linq; +using System.Numerics; +using System.Runtime.InteropServices; + +namespace Apache.Arrow +{ + /// <summary> + /// This is semi-optimised best attempt at converting to / from decimal and the buffers + /// </summary> + internal static class DecimalUtility + { + private static readonly BigInteger _maxDecimal = new BigInteger(decimal.MaxValue); + private static readonly BigInteger _minDecimal = new BigInteger(decimal.MinValue); + private static readonly ulong[] s_powersOfTen = + { + 1, 10, 100, 1000, 10000, 100000, 1000000, 10000000, 100000000, 1000000000, 10000000000, 100000000000, + 1000000000000, 10000000000000, 100000000000000, 1000000000000000, 10000000000000000, 100000000000000000, + 1000000000000000000, 10000000000000000000 + }; + private static int PowersOfTenLength => s_powersOfTen.Length - 1; + + public static decimal GetDecimal(in ArrowBuffer valueBuffer, int index, int scale, int byteWidth, + bool isUnsigned = false) + { + int startIndex = index * byteWidth; + ReadOnlySpan<byte> value = valueBuffer.Span.Slice(startIndex, byteWidth); + BigInteger integerValue; + +#if NETCOREAPP + integerValue = new BigInteger(value); +#else + integerValue = new BigInteger(value.ToArray()); +#endif + + if (integerValue > _maxDecimal || integerValue < _minDecimal) + { + BigInteger scaleBy = BigInteger.Pow(10, scale); + BigInteger integerPart = BigInteger.DivRem(integerValue, scaleBy, out BigInteger fractionalPart); + if (integerPart > _maxDecimal || integerPart < _minDecimal) // decimal overflow, not much we can do here - C# needs a BigDecimal + { + throw new OverflowException("Value: " + integerPart + " too big or too small to be represented as a decimal"); + } + return (decimal)integerPart + DivideByScale(fractionalPart, scale); + } + else + { + return DivideByScale(integerValue, scale); + } + } + + private static decimal DivideByScale(BigInteger integerValue, int scale) + { + decimal result = (decimal)integerValue; // this cast is safe here + int drop = scale; + while (drop > PowersOfTenLength) + { + result /= s_powersOfTen[PowersOfTenLength]; + drop -= PowersOfTenLength; + } + + result /= s_powersOfTen[drop]; + return result; + } + + public static void GetBytes(BigInteger integerValue, int byteWidth, ref Span<byte> bytes) + { + if (bytes.Length != byteWidth) + { + throw new OverflowException("ValueBuffer size not equal to " + byteWidth + " byte width: " + bytes.Length); + } + + Span<byte> integerBytes = integerValue.ToByteArray().AsSpan(); + if (integerBytes.Length > byteWidth) + { + throw new OverflowException("Decimal size greater than " + byteWidth + " bytes: " + integerBytes.Length); + } + + if (integerBytes.Length == byteWidth) + { + bytes = integerBytes; + return; + } + + if (integerValue.Sign == -1) + { + integerBytes.CopyTo(bytes); + for (int i = integerBytes.Length; i < byteWidth; i++) + { + bytes[i] = 255; + } + } + else + { + integerBytes.CopyTo(bytes); + } + } + + public static bool CheckPrecisionAndScale(decimal value, int precision, int scale, out BigInteger integerValue) + { + DecimalLayout layout = new DecimalLayout(value); // use in place of decimal.GetBits(value) to avoid an allocation + integerValue = new BigInteger(BitConverter.GetBytes(layout.Lo).Concat(BitConverter.GetBytes(layout.Mid)).Concat(BitConverter.GetBytes(layout.Hi)).ToArray()); Review comment: I've made another round of changes based on the feedback. These mostly combine CheckPrecisionAndScale and GetBytes - as you mention they need to be called together, and make some of the other suggestions / remove code which is not needed. [I have removed the DecimalLayout struct and tried to use netcore features where possibl - the exception being using the above for decimal.GetBits as a Span<int> which is only available on netcore 5.0 from what I can tell]. I've also added a few tests on the decimal utility, which is resulted in me making this public again, and also updated the docs. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected]
