使用R的xml2包解析CPLEX XML：提取变量名、索引与值

You’re really close! You’ve already targeted the <variables> node—now we just need to drill down to each individual <variable> element and pull out their attributes (name, index, value). Here’s how to do it smoothly:

Step-by-Step Code #

require(xml2)

# Your existing XML content and setup
xml_file <- '<?xml version = "1.0" encoding="UTF-8" standalone="yes"?>
<CPLEXSolution version="1.2">
<header problemName="Oil-blending.lp" objectiveValue="287750" solutionTypeValue="1" solutionTypeString="basic" solutionStatusValue="1" solutionStatusString="optimal" solutionMethodString="dual" primalFeasible="1" dualFeasible="1" simplexIterations="14" writeLevel="1"/>
<quality epRHS="1e-06" epOpt="1e-06" maxPrimalInfeas="0" maxDualInfeas="0" maxPrimalResidual="9.66338120633736e-13" maxDualResidual="7.105427357601e-15" maxX="7500" maxPi="57.25" maxSlack="4000" maxRedCost="40.9" kappa="83.7880434782609"/>
<linearConstraints>
<constraint name="ct_demand({&quot;Super&quot;})" index="0" status="LL" slack="0" dual="-20.8"/>
<constraint name="ct_demand({&quot;Regular&quot;})" index="1" status="LL" slack="0" dual="0.1"/>
<constraint name="ct_demand({&quot;Diesel&quot;})" index="2" status="LL" slack="0" dual="-40.8"/>
<constraint name="ct_capacity({&quot;Crude1&quot;})" index="3" status="LL" slack="0" dual="57.25"/>
<constraint name="ct_capacity({&quot;Crude2&quot;})" index="4" status="LL" slack="0" dual="20.9"/>
<constraint name="ct_capacity({&quot;Crude3&quot;})" index="5" status="BS" slack="1500" dual="0"/>
<constraint name="ct_total_max_prod" index="6" status="BS" slack="499.999999999997" dual="0"/>
<constraint name="ct_octane_min({&quot;Super&quot;})" index="7" status="BS" slack="-2000" dual="-0"/>
<constraint name="ct_octane_min({&quot;Regular&quot;})" index="8" status="LL" slack="0" dual="-1.77635683940025e-15"/>
<constraint name="ct_octane_min({&quot;Diesel&quot;})" index="9" status="BS" slack="-4000" dual="-0"/>
<constraint name="ct_lead_max({&quot;Super&quot;})" index="10" status="LL" slack="0" dual="30.9"/>
<constraint name="ct_lead_max({&quot;Regular&quot;})" index="11" status="LL" slack="0" dual="30.9"/>
<constraint name="ct_lead_max({&quot;Diesel&quot;})" index="12" status="LL" slack="0" dual="30.9"/>
</linearConstraints>
<variables>
<variable name="Blend({&quot;Crude1&quot;})({&quot;Super&quot;})" index="0" status="BS" value="2222.22222222222" reducedCost="-0"/>
<variable name="Blend({&quot;Crude2&quot;})({&quot;Super&quot;})" index="1" status="BS" value="444.444444444444" reducedCost="-0"/>
<variable name="Blend({&quot;Crude3&quot;})({&quot;Super&quot;})" index="2" status="BS" value="333.333333333333" reducedCost="-0"/>
<variable name="Blend({&quot;Crude1&quot;})({&quot;Regular&quot;})" index="3" status="BS" value="2111.11111111111" reducedCost="-0"/>
<variable name="Blend({&quot;Crude2&quot;})({&quot;Regular&quot;})" index="4" status="BS" value="4222.22222222222" reducedCost="-0"/>
<variable name="Blend({&quot;Crude3&quot;})({&quot;Regular&quot;})" index="5" status="BS" value="3166.66666666667" reducedCost="-0"/>
<variable name="Blend({&quot;Crude1&quot;})({&quot;Diesel&quot;})" index="6" status="BS" value="666.666666666667" reducedCost="-0"/>
<variable name="Blend({&quot;Crude2&quot;})({&quot;Diesel&quot;})" index="7" status="BS" value="333.333333333333" reducedCost="-0"/>
<variable name="Blend({&quot;Crude3&quot;})({&quot;Diesel&quot;})" index="8" status="LL" value="0" reducedCost="-7.105427357601e-15"/>
<variable name="Inventory({&quot;Super&quot;})" index="9" status="LL" value="0" reducedCost="-20.9"/>
<variable name="Inventory({&quot;Regular&quot;})" index="10" status="BS" value="7500" reducedCost="-0"/>
<variable name="Inventory({&quot;Diesel&quot;})" index="11" status="LL" value="0" reducedCost="-40.9"/>
<variable name="x13" index="12" status="UL" value="0" reducedCost="1"/>
</variables>
</CPLEXSolution>'

x <- read_xml(xml_file)

# Get every <variable> element inside the <variables> node
vars <- xml_find_all(x, "//variables/variable")

# Extract attributes and convert to usable data types
variable_df <- data.frame(
  Index = as.integer(xml_attr(vars, "index")),
  Name = xml_attr(vars, "name"),
  Value = as.numeric(xml_attr(vars, "value")),
  ReducedCost = as.numeric(xml_attr(vars, "reducedCost")), # Optional, if you need this metric
  stringsAsFactors = FALSE
)

# View the cleaned-up results
print(variable_df)

Key Explanations #

• Targeting individual variables: Instead of just selecting the parent <variables> node, we use the XPath //variables/variable to grab every child <variable> element.
• Extracting attributes: The xml_attr() function pulls the value of a specified attribute from each element in the vars list. We convert index to integer and numeric fields like value to numeric so they’re ready for analysis.
• Data frame organization: This structure makes it easy to filter, sort, or export the variable data for further work.

内容的提问来源于stack exchange，提问作者rwdvc