The impact of changes in incoming solar ultraviolet irradiance on stratospheric ozone forms an important part of the climate response to solar variability. To realistically simulate the climate response to solar variability using climate models, a minimum requirement is that they should include a solar cycle ozone component that has a realistic amplitude and structure, and which varies with season. For climate models that do not include interactive ozone chemistry, this component must be derived from observations and/or chemistry–climate model simulations and included in an externally prescribed ozone database that also includes the effects of all major external forcings. Part 1 of this two part study presents the solar-ozone responses in a number of updated satellite datasets for the period 1984–2004, including the Stratospheric Aerosol and Gas Experiment (SAGE) II version 6.2 and version 7.0 data, and the Solar Backscatter Ultraviolet Instrument (SBUV) version 8.0 and version 8.6 data. A number of combined datasets, which have extended SAGE II using more recent satellite measurements, are also analysed for the period 1984–2011. It is shown that SAGE II derived solar-ozone signals are sensitive to the independent temperature measurements used to convert ozone number density to mixing ratio units. A change in these temperature measurements in the recent SAGE II v7.0 data leads to substantial differences in the mixing ratio solar-ozone response compared to the previous v6.2, particularly in the tropical upper stratosphere. We also show that alternate satellite ozone datasets have issues (e.g., sparse spatial and temporal sampling, low vertical resolution, and shortness of measurement record), and that the methods of accounting for instrument offsets and drifts in merged satellite datasets can have a substantial impact on the solar cycle signal in ozone. For example, the magnitude of the solar-ozone response varies by around a factor of two across different versions of the SBUV VN8.6 record, which appears to be due to the methods used to combine the separate SBUV timeseries. These factors make it difficult to extract more than an annual-mean solar-ozone response from the available satellite observations. It is therefore unlikely that satellite ozone measurements alone can be applied to estimate the necessary solar cycle ozone component of the prescribed ozone database for future coupled model intercomparison projects (e.g., CMIP6).