求助：用AWK脚本过滤日志并提取指定参数到Shell变量

Hey there! Let's break this down step by step since you're new to shell scripting and awk—no worries, we'll get you sorted. First, let's start with a realistic example of what your log might look like (since you didn't share the exact format, I'll assume space-separated key-value pairs, which is common):

processor=gateway1 paymentType=credit amount=100.50 currency=USD network=VISA responseCode=00
processor=gateway2 paymentType=debit amount=75.00 currency=EUR network=MASTER responseCode=00
processor=gateway1 paymentType=credit amount=200.00 currency=GBP network=AMEX responseCode=01
processor=gateway3 paymentType=credit amount=50.00 currency=USD network=VISA responseCode=00

Step 1: Filter rows by processor and paymentType #

Your initial command only pulled lines with "amount", but we need to first narrow down to rows that match your target processor and payment type. Let's say you want processor=gateway1 and paymentType=credit—here's how to filter those with awk:

awk '$1 == "processor=gateway1" && $2 == "paymentType=credit"' testAPI.log

If the order of fields isn't fixed (e.g., processor might not always be first), a more robust approach checks every field for the keys:

awk '{
    match_proc=0; match_pay=0
    for (i=1; i<=NF; i++) {
        split($i, kv, "=")
        if (kv[1] == "processor" && kv[2] == "gateway1") match_proc=1
        if (kv[1] == "paymentType" && kv[2] == "credit") match_pay=1
    }
    if (match_proc && match_pay) print $0
}' testAPI.log

Step 2: Extract the required fields from filtered rows #

Now that we have the right rows, we need to pull out amount, currency, network, and responseCode. We'll extend the awk script to capture these values and print them in a clean format:

awk '{
    match_proc=0; match_pay=0
    amt=""; curr=""; net=""; rc=""
    # First check if the row matches our filters
    for (i=1; i<=NF; i++) {
        split($i, kv, "=")
        if (kv[1] == "processor" && kv[2] == "gateway1") match_proc=1
        if (kv[1] == "paymentType" && kv[2] == "credit") match_pay=1
        # Capture the values we need while we're looping
        if (kv[1] == "amount") amt=kv[2]
        if (kv[1] == "currency") curr=kv[2]
        if (kv[1] == "network") net=kv[2]
        if (kv[1] == "responseCode") rc=kv[2]
    }
    # Only print if the row matched our filters
    if (match_proc && match_pay) print amt, curr, net, rc
}' testAPI.log

Step 3: Store extracted values in shell variables #

To use these values in your Oracle insert command, we can read awk's output into shell variables. If you only expect one matching row, use read:

read amount currency network responseCode <<< $(awk '{
    match_proc=0; match_pay=0
    amt=""; curr=""; net=""; rc=""
    for (i=1; i<=NF; i++) {
        split($i, kv, "=")
        if (kv[1] == "processor" && kv[2] == "gateway1") match_proc=1
        if (kv[1] == "paymentType" && kv[2] == "credit") match_pay=1
        if (kv[1] == "amount") amt=kv[2]
        if (kv[1] == "currency") curr=kv[2]
        if (kv[1] == "network") net=kv[2]
        if (kv[1] == "responseCode") rc=kv[2]
    }
    if (match_proc && match_pay) print amt, curr, net, rc
}' testAPI.log)

# Verify the variables work
echo "Amount: $amount"
echo "Currency: $currency"
echo "Network: $network"
echo "Response Code: $responseCode"

If there are multiple matching rows, loop through each one and insert them into Oracle:

awk '{
    match_proc=0; match_pay=0
    amt=""; curr=""; net=""; rc=""
    for (i=1; i<=NF; i++) {
        split($i, kv, "=")
        if (kv[1] == "processor" && kv[2] == "gateway1") match_proc=1
        if (kv[1] == "paymentType" && kv[2] == "credit") match_pay=1
        if (kv[1] == "amount") amt=kv[2]
        if (kv[1] == "currency") curr=kv[2]
        if (kv[1] == "network") net=kv[2]
        if (kv[1] == "responseCode") rc=kv[2]
    }
    if (match_proc && match_pay) print amt, curr, net, rc
}' testAPI.log | while read amount currency network responseCode; do
    # Replace this with your actual Oracle insert command
    sqlplus your_username/your_password@your_database << EOF
    INSERT INTO your_table (amount, currency, network, response_code) 
    VALUES ($amount, '$currency', '$network', '$responseCode');
    COMMIT;
EOF
done

Quick notes for beginners: #

• Replace gateway1, credit, and the Oracle details with your actual values
• For string fields like currency, wrap them in single quotes in the SQL insert (since they're not numeric)
• If your log uses a different separator (not spaces), adjust the awk logic (e.g., change NF handling or split delimiter)

内容的提问来源于stack exchange，提问作者ChilBud