xts对象异常：单行列索引仅返回列名，blotter功能失效

Hey there, let's break down this frustrating xts indexing issue you're hitting with R 3.4.2 and those specific package versions. I've seen similar quirks with older R/xts combinations, so here's what's going on and how to fix it:

Key Issue Recap #

First, to make sure we're on the same page:

• Single-row indexing like IBM[1,] or IBM[21,] only spits out column names, no actual price data
• Multi-row indexing (e.g., IBM[1:2,]) works as expected
• The blotter package's amzn_test demo fails
• You've confirmed the problem reproduces in a minimal environment with just xts and the IBM data (your DUMMY package), so it's not a conflict with other packages

Why This Is Happening #

This boils down to two main factors:

1. Version incompatibility: R 3.4.2 is quite old (released back in 2017), and xts_0.10-2.1 has subtle bugs or compatibility gaps with this older R version that break single-row subsetting logic.
2. xts's older single-row handling: Early xts versions had known issues with preserving object structure when subsetting a single row—sometimes it would coerce the result to a vector instead of retaining the xts format, which leads to the weird "only column names" output when using [,] indexing.

Fixes & Workarounds #

1. Quick Workaround for Single-Row Indexing

Force xts to keep the object structure by adding drop=FALSE to your indexing call. This prevents the coercion to a vector that's causing the issue:

# This should return the full row data, not just column names
IBM[1, , drop = FALSE]
IBM[21, , drop = FALSE]

Alternatively, use the window() function, which is designed for time-series subsetting and handles single periods more reliably in older xts versions:

window(IBM, start = index(IBM)[1], end = index(IBM)[1])

2. Upgrade Packages (If Possible)

Since R 3.4.2 is end-of-life, upgrading to a newer R version (preferably 4.x) and updating xts/blotter to their latest stable releases will almost certainly resolve this issue long-term. If you can't upgrade R right now, install the latest compatible versions for R 3.4.2 from CRAN's archive:

# Install compatible xts version
install.packages("https://cran.r-project.org/src/contrib/Archive/xts/xts_0.12.1.tar.gz", repos = NULL, type = "source")
# Install compatible blotter version
install.packages("https://cran.r-project.org/src/contrib/Archive/blotter/blotter_0.14.4.tar.gz", repos = NULL, type = "source")

Note: You might need to update dependencies like zoo first if they're out of date.

3. Check & Adjust the IBM Data Structure

Verify if the built-in IBM data is a single-column xts object (this can trigger the indexing bug). Use str(IBM) to inspect:

str(IBM)

If it's a single-column xts, you can temporarily convert it to a data frame for indexing, then convert back to xts if needed:

ibm_df <- as.data.frame(IBM)
ibm_df[1, ] # This will return the row data
# Convert back to xts if required
ibm_xts <- xts(ibm_df, order.by = index(IBM))

4. Fix the amzn_test Demo

The amzn_test failure is almost certainly tied to the same xts indexing issue. Look for single-row indexing calls in the demo code and add the drop=FALSE workaround. After updating xts/blotter, re-run the demo to confirm it works.

内容的提问来源于stack exchange，提问作者Sean Sinykin