All Downloads are FREE. Search and download functionalities are using the official Maven repository.

com.amazonaws.services.kinesisfirehose.model.OrcSerDe Maven / Gradle / Ivy

Go to download

The AWS Java SDK for Amazon Kinesis module holds the client classes that are used for communicating with Amazon Kinesis Service

There is a newer version: 1.12.772
Show newest version
/*
 * Copyright 2017-2022 Amazon.com, Inc. or its affiliates. All Rights Reserved.
 * 
 * Licensed under the Apache License, Version 2.0 (the "License"). You may not use this file except in compliance with
 * the License. A copy of the License is located at
 * 
 * http://aws.amazon.com/apache2.0
 * 
 * or in the "license" file accompanying this file. This file 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 com.amazonaws.services.kinesisfirehose.model;

import java.io.Serializable;
import javax.annotation.Generated;
import com.amazonaws.protocol.StructuredPojo;
import com.amazonaws.protocol.ProtocolMarshaller;

/**
 * 

* A serializer to use for converting data to the ORC format before storing it in Amazon S3. For more information, see * Apache ORC. *

* * @see AWS API * Documentation */ @Generated("com.amazonaws:aws-java-sdk-code-generator") public class OrcSerDe implements Serializable, Cloneable, StructuredPojo { /** *

* The number of bytes in each stripe. The default is 64 MiB and the minimum is 8 MiB. *

*/ private Integer stripeSizeBytes; /** *

* The Hadoop Distributed File System (HDFS) block size. This is useful if you intend to copy the data from Amazon * S3 to HDFS before querying. The default is 256 MiB and the minimum is 64 MiB. Kinesis Data Firehose uses this * value for padding calculations. *

*/ private Integer blockSizeBytes; /** *

* The number of rows between index entries. The default is 10,000 and the minimum is 1,000. *

*/ private Integer rowIndexStride; /** *

* Set this to true to indicate that you want stripes to be padded to the HDFS block boundaries. This * is useful if you intend to copy the data from Amazon S3 to HDFS before querying. The default is * false. *

*/ private Boolean enablePadding; /** *

* A number between 0 and 1 that defines the tolerance for block padding as a decimal fraction of stripe size. The * default value is 0.05, which means 5 percent of stripe size. *

*

* For the default values of 64 MiB ORC stripes and 256 MiB HDFS blocks, the default block padding tolerance of 5 * percent reserves a maximum of 3.2 MiB for padding within the 256 MiB block. In such a case, if the available size * within the block is more than 3.2 MiB, a new, smaller stripe is inserted to fit within that space. This ensures * that no stripe crosses block boundaries and causes remote reads within a node-local task. *

*

* Kinesis Data Firehose ignores this parameter when OrcSerDe$EnablePadding is false. *

*/ private Double paddingTolerance; /** *

* The compression code to use over data blocks. The default is SNAPPY. *

*/ private String compression; /** *

* The column names for which you want Kinesis Data Firehose to create bloom filters. The default is * null. *

*/ private java.util.List bloomFilterColumns; /** *

* The Bloom filter false positive probability (FPP). The lower the FPP, the bigger the Bloom filter. The default * value is 0.05, the minimum is 0, and the maximum is 1. *

*/ private Double bloomFilterFalsePositiveProbability; /** *

* Represents the fraction of the total number of non-null rows. To turn off dictionary encoding, set this fraction * to a number that is less than the number of distinct keys in a dictionary. To always use dictionary encoding, set * this threshold to 1. *

*/ private Double dictionaryKeyThreshold; /** *

* The version of the file to write. The possible values are V0_11 and V0_12. The default * is V0_12. *

*/ private String formatVersion; /** *

* The number of bytes in each stripe. The default is 64 MiB and the minimum is 8 MiB. *

* * @param stripeSizeBytes * The number of bytes in each stripe. The default is 64 MiB and the minimum is 8 MiB. */ public void setStripeSizeBytes(Integer stripeSizeBytes) { this.stripeSizeBytes = stripeSizeBytes; } /** *

* The number of bytes in each stripe. The default is 64 MiB and the minimum is 8 MiB. *

* * @return The number of bytes in each stripe. The default is 64 MiB and the minimum is 8 MiB. */ public Integer getStripeSizeBytes() { return this.stripeSizeBytes; } /** *

* The number of bytes in each stripe. The default is 64 MiB and the minimum is 8 MiB. *

* * @param stripeSizeBytes * The number of bytes in each stripe. The default is 64 MiB and the minimum is 8 MiB. * @return Returns a reference to this object so that method calls can be chained together. */ public OrcSerDe withStripeSizeBytes(Integer stripeSizeBytes) { setStripeSizeBytes(stripeSizeBytes); return this; } /** *

* The Hadoop Distributed File System (HDFS) block size. This is useful if you intend to copy the data from Amazon * S3 to HDFS before querying. The default is 256 MiB and the minimum is 64 MiB. Kinesis Data Firehose uses this * value for padding calculations. *

* * @param blockSizeBytes * The Hadoop Distributed File System (HDFS) block size. This is useful if you intend to copy the data from * Amazon S3 to HDFS before querying. The default is 256 MiB and the minimum is 64 MiB. Kinesis Data Firehose * uses this value for padding calculations. */ public void setBlockSizeBytes(Integer blockSizeBytes) { this.blockSizeBytes = blockSizeBytes; } /** *

* The Hadoop Distributed File System (HDFS) block size. This is useful if you intend to copy the data from Amazon * S3 to HDFS before querying. The default is 256 MiB and the minimum is 64 MiB. Kinesis Data Firehose uses this * value for padding calculations. *

* * @return The Hadoop Distributed File System (HDFS) block size. This is useful if you intend to copy the data from * Amazon S3 to HDFS before querying. The default is 256 MiB and the minimum is 64 MiB. Kinesis Data * Firehose uses this value for padding calculations. */ public Integer getBlockSizeBytes() { return this.blockSizeBytes; } /** *

* The Hadoop Distributed File System (HDFS) block size. This is useful if you intend to copy the data from Amazon * S3 to HDFS before querying. The default is 256 MiB and the minimum is 64 MiB. Kinesis Data Firehose uses this * value for padding calculations. *

* * @param blockSizeBytes * The Hadoop Distributed File System (HDFS) block size. This is useful if you intend to copy the data from * Amazon S3 to HDFS before querying. The default is 256 MiB and the minimum is 64 MiB. Kinesis Data Firehose * uses this value for padding calculations. * @return Returns a reference to this object so that method calls can be chained together. */ public OrcSerDe withBlockSizeBytes(Integer blockSizeBytes) { setBlockSizeBytes(blockSizeBytes); return this; } /** *

* The number of rows between index entries. The default is 10,000 and the minimum is 1,000. *

* * @param rowIndexStride * The number of rows between index entries. The default is 10,000 and the minimum is 1,000. */ public void setRowIndexStride(Integer rowIndexStride) { this.rowIndexStride = rowIndexStride; } /** *

* The number of rows between index entries. The default is 10,000 and the minimum is 1,000. *

* * @return The number of rows between index entries. The default is 10,000 and the minimum is 1,000. */ public Integer getRowIndexStride() { return this.rowIndexStride; } /** *

* The number of rows between index entries. The default is 10,000 and the minimum is 1,000. *

* * @param rowIndexStride * The number of rows between index entries. The default is 10,000 and the minimum is 1,000. * @return Returns a reference to this object so that method calls can be chained together. */ public OrcSerDe withRowIndexStride(Integer rowIndexStride) { setRowIndexStride(rowIndexStride); return this; } /** *

* Set this to true to indicate that you want stripes to be padded to the HDFS block boundaries. This * is useful if you intend to copy the data from Amazon S3 to HDFS before querying. The default is * false. *

* * @param enablePadding * Set this to true to indicate that you want stripes to be padded to the HDFS block boundaries. * This is useful if you intend to copy the data from Amazon S3 to HDFS before querying. The default is * false. */ public void setEnablePadding(Boolean enablePadding) { this.enablePadding = enablePadding; } /** *

* Set this to true to indicate that you want stripes to be padded to the HDFS block boundaries. This * is useful if you intend to copy the data from Amazon S3 to HDFS before querying. The default is * false. *

* * @return Set this to true to indicate that you want stripes to be padded to the HDFS block * boundaries. This is useful if you intend to copy the data from Amazon S3 to HDFS before querying. The * default is false. */ public Boolean getEnablePadding() { return this.enablePadding; } /** *

* Set this to true to indicate that you want stripes to be padded to the HDFS block boundaries. This * is useful if you intend to copy the data from Amazon S3 to HDFS before querying. The default is * false. *

* * @param enablePadding * Set this to true to indicate that you want stripes to be padded to the HDFS block boundaries. * This is useful if you intend to copy the data from Amazon S3 to HDFS before querying. The default is * false. * @return Returns a reference to this object so that method calls can be chained together. */ public OrcSerDe withEnablePadding(Boolean enablePadding) { setEnablePadding(enablePadding); return this; } /** *

* Set this to true to indicate that you want stripes to be padded to the HDFS block boundaries. This * is useful if you intend to copy the data from Amazon S3 to HDFS before querying. The default is * false. *

* * @return Set this to true to indicate that you want stripes to be padded to the HDFS block * boundaries. This is useful if you intend to copy the data from Amazon S3 to HDFS before querying. The * default is false. */ public Boolean isEnablePadding() { return this.enablePadding; } /** *

* A number between 0 and 1 that defines the tolerance for block padding as a decimal fraction of stripe size. The * default value is 0.05, which means 5 percent of stripe size. *

*

* For the default values of 64 MiB ORC stripes and 256 MiB HDFS blocks, the default block padding tolerance of 5 * percent reserves a maximum of 3.2 MiB for padding within the 256 MiB block. In such a case, if the available size * within the block is more than 3.2 MiB, a new, smaller stripe is inserted to fit within that space. This ensures * that no stripe crosses block boundaries and causes remote reads within a node-local task. *

*

* Kinesis Data Firehose ignores this parameter when OrcSerDe$EnablePadding is false. *

* * @param paddingTolerance * A number between 0 and 1 that defines the tolerance for block padding as a decimal fraction of stripe * size. The default value is 0.05, which means 5 percent of stripe size.

*

* For the default values of 64 MiB ORC stripes and 256 MiB HDFS blocks, the default block padding tolerance * of 5 percent reserves a maximum of 3.2 MiB for padding within the 256 MiB block. In such a case, if the * available size within the block is more than 3.2 MiB, a new, smaller stripe is inserted to fit within that * space. This ensures that no stripe crosses block boundaries and causes remote reads within a node-local * task. *

*

* Kinesis Data Firehose ignores this parameter when OrcSerDe$EnablePadding is false. */ public void setPaddingTolerance(Double paddingTolerance) { this.paddingTolerance = paddingTolerance; } /** *

* A number between 0 and 1 that defines the tolerance for block padding as a decimal fraction of stripe size. The * default value is 0.05, which means 5 percent of stripe size. *

*

* For the default values of 64 MiB ORC stripes and 256 MiB HDFS blocks, the default block padding tolerance of 5 * percent reserves a maximum of 3.2 MiB for padding within the 256 MiB block. In such a case, if the available size * within the block is more than 3.2 MiB, a new, smaller stripe is inserted to fit within that space. This ensures * that no stripe crosses block boundaries and causes remote reads within a node-local task. *

*

* Kinesis Data Firehose ignores this parameter when OrcSerDe$EnablePadding is false. *

* * @return A number between 0 and 1 that defines the tolerance for block padding as a decimal fraction of stripe * size. The default value is 0.05, which means 5 percent of stripe size.

*

* For the default values of 64 MiB ORC stripes and 256 MiB HDFS blocks, the default block padding tolerance * of 5 percent reserves a maximum of 3.2 MiB for padding within the 256 MiB block. In such a case, if the * available size within the block is more than 3.2 MiB, a new, smaller stripe is inserted to fit within * that space. This ensures that no stripe crosses block boundaries and causes remote reads within a * node-local task. *

*

* Kinesis Data Firehose ignores this parameter when OrcSerDe$EnablePadding is false. */ public Double getPaddingTolerance() { return this.paddingTolerance; } /** *

* A number between 0 and 1 that defines the tolerance for block padding as a decimal fraction of stripe size. The * default value is 0.05, which means 5 percent of stripe size. *

*

* For the default values of 64 MiB ORC stripes and 256 MiB HDFS blocks, the default block padding tolerance of 5 * percent reserves a maximum of 3.2 MiB for padding within the 256 MiB block. In such a case, if the available size * within the block is more than 3.2 MiB, a new, smaller stripe is inserted to fit within that space. This ensures * that no stripe crosses block boundaries and causes remote reads within a node-local task. *

*

* Kinesis Data Firehose ignores this parameter when OrcSerDe$EnablePadding is false. *

* * @param paddingTolerance * A number between 0 and 1 that defines the tolerance for block padding as a decimal fraction of stripe * size. The default value is 0.05, which means 5 percent of stripe size.

*

* For the default values of 64 MiB ORC stripes and 256 MiB HDFS blocks, the default block padding tolerance * of 5 percent reserves a maximum of 3.2 MiB for padding within the 256 MiB block. In such a case, if the * available size within the block is more than 3.2 MiB, a new, smaller stripe is inserted to fit within that * space. This ensures that no stripe crosses block boundaries and causes remote reads within a node-local * task. *

*

* Kinesis Data Firehose ignores this parameter when OrcSerDe$EnablePadding is false. * @return Returns a reference to this object so that method calls can be chained together. */ public OrcSerDe withPaddingTolerance(Double paddingTolerance) { setPaddingTolerance(paddingTolerance); return this; } /** *

* The compression code to use over data blocks. The default is SNAPPY. *

* * @param compression * The compression code to use over data blocks. The default is SNAPPY. * @see OrcCompression */ public void setCompression(String compression) { this.compression = compression; } /** *

* The compression code to use over data blocks. The default is SNAPPY. *

* * @return The compression code to use over data blocks. The default is SNAPPY. * @see OrcCompression */ public String getCompression() { return this.compression; } /** *

* The compression code to use over data blocks. The default is SNAPPY. *

* * @param compression * The compression code to use over data blocks. The default is SNAPPY. * @return Returns a reference to this object so that method calls can be chained together. * @see OrcCompression */ public OrcSerDe withCompression(String compression) { setCompression(compression); return this; } /** *

* The compression code to use over data blocks. The default is SNAPPY. *

* * @param compression * The compression code to use over data blocks. The default is SNAPPY. * @return Returns a reference to this object so that method calls can be chained together. * @see OrcCompression */ public OrcSerDe withCompression(OrcCompression compression) { this.compression = compression.toString(); return this; } /** *

* The column names for which you want Kinesis Data Firehose to create bloom filters. The default is * null. *

* * @return The column names for which you want Kinesis Data Firehose to create bloom filters. The default is * null. */ public java.util.List getBloomFilterColumns() { return bloomFilterColumns; } /** *

* The column names for which you want Kinesis Data Firehose to create bloom filters. The default is * null. *

* * @param bloomFilterColumns * The column names for which you want Kinesis Data Firehose to create bloom filters. The default is * null. */ public void setBloomFilterColumns(java.util.Collection bloomFilterColumns) { if (bloomFilterColumns == null) { this.bloomFilterColumns = null; return; } this.bloomFilterColumns = new java.util.ArrayList(bloomFilterColumns); } /** *

* The column names for which you want Kinesis Data Firehose to create bloom filters. The default is * null. *

*

* NOTE: This method appends the values to the existing list (if any). Use * {@link #setBloomFilterColumns(java.util.Collection)} or {@link #withBloomFilterColumns(java.util.Collection)} if * you want to override the existing values. *

* * @param bloomFilterColumns * The column names for which you want Kinesis Data Firehose to create bloom filters. The default is * null. * @return Returns a reference to this object so that method calls can be chained together. */ public OrcSerDe withBloomFilterColumns(String... bloomFilterColumns) { if (this.bloomFilterColumns == null) { setBloomFilterColumns(new java.util.ArrayList(bloomFilterColumns.length)); } for (String ele : bloomFilterColumns) { this.bloomFilterColumns.add(ele); } return this; } /** *

* The column names for which you want Kinesis Data Firehose to create bloom filters. The default is * null. *

* * @param bloomFilterColumns * The column names for which you want Kinesis Data Firehose to create bloom filters. The default is * null. * @return Returns a reference to this object so that method calls can be chained together. */ public OrcSerDe withBloomFilterColumns(java.util.Collection bloomFilterColumns) { setBloomFilterColumns(bloomFilterColumns); return this; } /** *

* The Bloom filter false positive probability (FPP). The lower the FPP, the bigger the Bloom filter. The default * value is 0.05, the minimum is 0, and the maximum is 1. *

* * @param bloomFilterFalsePositiveProbability * The Bloom filter false positive probability (FPP). The lower the FPP, the bigger the Bloom filter. The * default value is 0.05, the minimum is 0, and the maximum is 1. */ public void setBloomFilterFalsePositiveProbability(Double bloomFilterFalsePositiveProbability) { this.bloomFilterFalsePositiveProbability = bloomFilterFalsePositiveProbability; } /** *

* The Bloom filter false positive probability (FPP). The lower the FPP, the bigger the Bloom filter. The default * value is 0.05, the minimum is 0, and the maximum is 1. *

* * @return The Bloom filter false positive probability (FPP). The lower the FPP, the bigger the Bloom filter. The * default value is 0.05, the minimum is 0, and the maximum is 1. */ public Double getBloomFilterFalsePositiveProbability() { return this.bloomFilterFalsePositiveProbability; } /** *

* The Bloom filter false positive probability (FPP). The lower the FPP, the bigger the Bloom filter. The default * value is 0.05, the minimum is 0, and the maximum is 1. *

* * @param bloomFilterFalsePositiveProbability * The Bloom filter false positive probability (FPP). The lower the FPP, the bigger the Bloom filter. The * default value is 0.05, the minimum is 0, and the maximum is 1. * @return Returns a reference to this object so that method calls can be chained together. */ public OrcSerDe withBloomFilterFalsePositiveProbability(Double bloomFilterFalsePositiveProbability) { setBloomFilterFalsePositiveProbability(bloomFilterFalsePositiveProbability); return this; } /** *

* Represents the fraction of the total number of non-null rows. To turn off dictionary encoding, set this fraction * to a number that is less than the number of distinct keys in a dictionary. To always use dictionary encoding, set * this threshold to 1. *

* * @param dictionaryKeyThreshold * Represents the fraction of the total number of non-null rows. To turn off dictionary encoding, set this * fraction to a number that is less than the number of distinct keys in a dictionary. To always use * dictionary encoding, set this threshold to 1. */ public void setDictionaryKeyThreshold(Double dictionaryKeyThreshold) { this.dictionaryKeyThreshold = dictionaryKeyThreshold; } /** *

* Represents the fraction of the total number of non-null rows. To turn off dictionary encoding, set this fraction * to a number that is less than the number of distinct keys in a dictionary. To always use dictionary encoding, set * this threshold to 1. *

* * @return Represents the fraction of the total number of non-null rows. To turn off dictionary encoding, set this * fraction to a number that is less than the number of distinct keys in a dictionary. To always use * dictionary encoding, set this threshold to 1. */ public Double getDictionaryKeyThreshold() { return this.dictionaryKeyThreshold; } /** *

* Represents the fraction of the total number of non-null rows. To turn off dictionary encoding, set this fraction * to a number that is less than the number of distinct keys in a dictionary. To always use dictionary encoding, set * this threshold to 1. *

* * @param dictionaryKeyThreshold * Represents the fraction of the total number of non-null rows. To turn off dictionary encoding, set this * fraction to a number that is less than the number of distinct keys in a dictionary. To always use * dictionary encoding, set this threshold to 1. * @return Returns a reference to this object so that method calls can be chained together. */ public OrcSerDe withDictionaryKeyThreshold(Double dictionaryKeyThreshold) { setDictionaryKeyThreshold(dictionaryKeyThreshold); return this; } /** *

* The version of the file to write. The possible values are V0_11 and V0_12. The default * is V0_12. *

* * @param formatVersion * The version of the file to write. The possible values are V0_11 and V0_12. The * default is V0_12. * @see OrcFormatVersion */ public void setFormatVersion(String formatVersion) { this.formatVersion = formatVersion; } /** *

* The version of the file to write. The possible values are V0_11 and V0_12. The default * is V0_12. *

* * @return The version of the file to write. The possible values are V0_11 and V0_12. The * default is V0_12. * @see OrcFormatVersion */ public String getFormatVersion() { return this.formatVersion; } /** *

* The version of the file to write. The possible values are V0_11 and V0_12. The default * is V0_12. *

* * @param formatVersion * The version of the file to write. The possible values are V0_11 and V0_12. The * default is V0_12. * @return Returns a reference to this object so that method calls can be chained together. * @see OrcFormatVersion */ public OrcSerDe withFormatVersion(String formatVersion) { setFormatVersion(formatVersion); return this; } /** *

* The version of the file to write. The possible values are V0_11 and V0_12. The default * is V0_12. *

* * @param formatVersion * The version of the file to write. The possible values are V0_11 and V0_12. The * default is V0_12. * @return Returns a reference to this object so that method calls can be chained together. * @see OrcFormatVersion */ public OrcSerDe withFormatVersion(OrcFormatVersion formatVersion) { this.formatVersion = formatVersion.toString(); return this; } /** * Returns a string representation of this object. This is useful for testing and debugging. Sensitive data will be * redacted from this string using a placeholder value. * * @return A string representation of this object. * * @see java.lang.Object#toString() */ @Override public String toString() { StringBuilder sb = new StringBuilder(); sb.append("{"); if (getStripeSizeBytes() != null) sb.append("StripeSizeBytes: ").append(getStripeSizeBytes()).append(","); if (getBlockSizeBytes() != null) sb.append("BlockSizeBytes: ").append(getBlockSizeBytes()).append(","); if (getRowIndexStride() != null) sb.append("RowIndexStride: ").append(getRowIndexStride()).append(","); if (getEnablePadding() != null) sb.append("EnablePadding: ").append(getEnablePadding()).append(","); if (getPaddingTolerance() != null) sb.append("PaddingTolerance: ").append(getPaddingTolerance()).append(","); if (getCompression() != null) sb.append("Compression: ").append(getCompression()).append(","); if (getBloomFilterColumns() != null) sb.append("BloomFilterColumns: ").append(getBloomFilterColumns()).append(","); if (getBloomFilterFalsePositiveProbability() != null) sb.append("BloomFilterFalsePositiveProbability: ").append(getBloomFilterFalsePositiveProbability()).append(","); if (getDictionaryKeyThreshold() != null) sb.append("DictionaryKeyThreshold: ").append(getDictionaryKeyThreshold()).append(","); if (getFormatVersion() != null) sb.append("FormatVersion: ").append(getFormatVersion()); sb.append("}"); return sb.toString(); } @Override public boolean equals(Object obj) { if (this == obj) return true; if (obj == null) return false; if (obj instanceof OrcSerDe == false) return false; OrcSerDe other = (OrcSerDe) obj; if (other.getStripeSizeBytes() == null ^ this.getStripeSizeBytes() == null) return false; if (other.getStripeSizeBytes() != null && other.getStripeSizeBytes().equals(this.getStripeSizeBytes()) == false) return false; if (other.getBlockSizeBytes() == null ^ this.getBlockSizeBytes() == null) return false; if (other.getBlockSizeBytes() != null && other.getBlockSizeBytes().equals(this.getBlockSizeBytes()) == false) return false; if (other.getRowIndexStride() == null ^ this.getRowIndexStride() == null) return false; if (other.getRowIndexStride() != null && other.getRowIndexStride().equals(this.getRowIndexStride()) == false) return false; if (other.getEnablePadding() == null ^ this.getEnablePadding() == null) return false; if (other.getEnablePadding() != null && other.getEnablePadding().equals(this.getEnablePadding()) == false) return false; if (other.getPaddingTolerance() == null ^ this.getPaddingTolerance() == null) return false; if (other.getPaddingTolerance() != null && other.getPaddingTolerance().equals(this.getPaddingTolerance()) == false) return false; if (other.getCompression() == null ^ this.getCompression() == null) return false; if (other.getCompression() != null && other.getCompression().equals(this.getCompression()) == false) return false; if (other.getBloomFilterColumns() == null ^ this.getBloomFilterColumns() == null) return false; if (other.getBloomFilterColumns() != null && other.getBloomFilterColumns().equals(this.getBloomFilterColumns()) == false) return false; if (other.getBloomFilterFalsePositiveProbability() == null ^ this.getBloomFilterFalsePositiveProbability() == null) return false; if (other.getBloomFilterFalsePositiveProbability() != null && other.getBloomFilterFalsePositiveProbability().equals(this.getBloomFilterFalsePositiveProbability()) == false) return false; if (other.getDictionaryKeyThreshold() == null ^ this.getDictionaryKeyThreshold() == null) return false; if (other.getDictionaryKeyThreshold() != null && other.getDictionaryKeyThreshold().equals(this.getDictionaryKeyThreshold()) == false) return false; if (other.getFormatVersion() == null ^ this.getFormatVersion() == null) return false; if (other.getFormatVersion() != null && other.getFormatVersion().equals(this.getFormatVersion()) == false) return false; return true; } @Override public int hashCode() { final int prime = 31; int hashCode = 1; hashCode = prime * hashCode + ((getStripeSizeBytes() == null) ? 0 : getStripeSizeBytes().hashCode()); hashCode = prime * hashCode + ((getBlockSizeBytes() == null) ? 0 : getBlockSizeBytes().hashCode()); hashCode = prime * hashCode + ((getRowIndexStride() == null) ? 0 : getRowIndexStride().hashCode()); hashCode = prime * hashCode + ((getEnablePadding() == null) ? 0 : getEnablePadding().hashCode()); hashCode = prime * hashCode + ((getPaddingTolerance() == null) ? 0 : getPaddingTolerance().hashCode()); hashCode = prime * hashCode + ((getCompression() == null) ? 0 : getCompression().hashCode()); hashCode = prime * hashCode + ((getBloomFilterColumns() == null) ? 0 : getBloomFilterColumns().hashCode()); hashCode = prime * hashCode + ((getBloomFilterFalsePositiveProbability() == null) ? 0 : getBloomFilterFalsePositiveProbability().hashCode()); hashCode = prime * hashCode + ((getDictionaryKeyThreshold() == null) ? 0 : getDictionaryKeyThreshold().hashCode()); hashCode = prime * hashCode + ((getFormatVersion() == null) ? 0 : getFormatVersion().hashCode()); return hashCode; } @Override public OrcSerDe clone() { try { return (OrcSerDe) super.clone(); } catch (CloneNotSupportedException e) { throw new IllegalStateException("Got a CloneNotSupportedException from Object.clone() " + "even though we're Cloneable!", e); } } @com.amazonaws.annotation.SdkInternalApi @Override public void marshall(ProtocolMarshaller protocolMarshaller) { com.amazonaws.services.kinesisfirehose.model.transform.OrcSerDeMarshaller.getInstance().marshall(this, protocolMarshaller); } }




© 2015 - 2024 Weber Informatics LLC | Privacy Policy